Vehicle product analysis decision method, device and equipment

Abstract

The invention discloses a vehicle product analysis decision method, device and equipment. The method comprises the following steps: calling a pre-constructed knowledge search tool based on a pre-constructed large model for vehicle product analysis and decision, searching first vehicle use scene information, first vehicle product project information and first vehicle product constraint information associated with the vehicle product demand analysis data from a pre-constructed vehicle knowledge base; based on the first vehicle product constraint information, determining a plurality of vehicle product decision factors matched with the vehicle product demand analysis data, calling a pre-constructed constraint and risk assessment tool, and based on the plurality of vehicle product decision factors, determining a vehicle product risk assessment result; and generating a vehicle product analysis decision result corresponding to the vehicle product demand analysis data based on the first vehicle use scene information, the first vehicle product project information and the vehicle product risk assessment result. Therefore, the accuracy of vehicle product analysis decision can be improved.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a method, apparatus and equipment for vehicle product analysis and decision-making. Background Technology

[0002] Currently, user needs identification and product analysis decision-making primarily rely on user surveys, questionnaires, and scenario analysis to collect user requirements for product feature decisions. This approach is lengthy and complex, and issues such as data sample coverage can frequently lead to deviations in analysis results. Ultimately, this results in a significant investment of effort in requirements identification and product definition without achieving satisfactory results, or even in products that contradict user needs, leading to product development failure.

[0003] To address the aforementioned issues, product personnel have incorporated tools such as scenario simulation, value assessment, the Kano model (a theoretical framework for prioritizing and classifying user needs), and Quality Function Deployment (QFD) into the requirements identification and product definition processes to assist in analysis and decision-making, thereby reducing deviations in product definition. However, the discovery of requirements and product decisions still rely heavily on the experience and intuition of product personnel, combined with market trends and user feedback. Significant challenges remain in the discovery, identification, and decision-making regarding deep and broadly interconnected requirements.

[0004] Therefore, improving the accuracy of vehicle product analysis and decision-making has become an urgent problem to be solved. Summary of the Invention

[0005] This application provides a vehicle product analysis and decision-making method, apparatus, and equipment, which can improve the accuracy of vehicle product analysis and decision-making.

[0006] In a first aspect, embodiments of this application provide a vehicle product analysis and decision-making method, the method comprising:

[0007] Obtain vehicle product demand analysis data input by the target object;

[0008] Based on a pre-built large model for vehicle product analysis and decision-making, a pre-built knowledge search tool is invoked to search the pre-built vehicle knowledge base for first vehicle usage scenario information, first vehicle product project information, and first vehicle product constraint information related to vehicle product demand analysis data. The vehicle knowledge base includes vehicle usage scenario data for reconstructing real demand scenarios through multi-dimensional vehicle usage scenarios, vehicle product project data for aligning demand analysis with actual goals, and vehicle product constraint data for verifying the feasibility of demands. The first vehicle usage scenario information is obtained from the vehicle usage scenario data, the first vehicle product project information is obtained from the vehicle product project data, and the first vehicle product constraint information is obtained from the vehicle product constraint data.

[0009] Based on the first vehicle product constraint information, multiple vehicle product decision factors that match the vehicle product demand analysis data are identified, and a pre-built constraint and risk assessment tool is invoked to determine the vehicle product risk assessment result corresponding to the vehicle product demand analysis data based on the multiple vehicle product decision factors.

[0010] Based on the first vehicle usage scenario information, the first vehicle product project information, and the vehicle product risk assessment results, the vehicle product analysis decision results corresponding to the vehicle product demand analysis data are generated.

[0011] In one embodiment, the method further includes: invoking a pre-built network search tool to search the entire network for second vehicle usage scenario information, second vehicle product project information, and second vehicle product constraint information associated with the vehicle product demand analysis data; determining multiple vehicle product decision factors matching the vehicle product demand analysis data based on the first vehicle product constraint information, including: extracting multiple vehicle product decision factors matching the vehicle product demand analysis data from the first vehicle product constraint information and the second vehicle product constraint information; and generating a vehicle product analysis decision result corresponding to the vehicle product demand analysis data based on the first vehicle usage scenario information, the first vehicle product project information, and the vehicle product risk assessment result, including: generating a vehicle product analysis decision result corresponding to the vehicle product demand analysis data based on the first vehicle usage scenario information, the first vehicle product project information, the second vehicle usage scenario information, the second vehicle product project information, and the vehicle product risk assessment result.

[0012] In one embodiment, the vehicle product risk assessment result includes the target vehicle product risk assessment score corresponding to the vehicle product demand analysis data; determining the vehicle product risk assessment result corresponding to the vehicle product demand analysis data based on multiple vehicle product decision factors includes: obtaining the target weight and vehicle product risk assessment score corresponding to each vehicle product decision factor among the multiple vehicle product decision factors; the target weight is preset or determined based on a pre-determined weight allocation strategy; and determining the vehicle product risk assessment result corresponding to the vehicle product demand analysis data based on the target weight and vehicle product risk assessment score corresponding to each vehicle product decision factor.

[0013] In one embodiment, the method further includes: determining the category to which the vehicle product decision factor belongs and the decision level to which the vehicle product decision factor belongs based on the attributes corresponding to the vehicle product decision factor; determining the initial weight of the vehicle product decision factor based on the decision level; and determining the weight correction coefficient corresponding to the vehicle product decision factor based on the category to which the vehicle product decision factor belongs and the vehicle product risk assessment score corresponding to the vehicle product decision factor; and correcting the initial weight based on the weight correction coefficient to obtain the target weight corresponding to the vehicle product decision factor.

[0014] In one embodiment, the multiple vehicle product decision factors include at least two of a first vehicle product decision factor, a second vehicle product decision factor, a third vehicle product decision factor, a fourth vehicle product decision factor, a fifth vehicle product decision factor, and a sixth vehicle product decision factor. The first vehicle product decision factor is used to conduct a risk assessment of the vehicle product demand analysis data from a first dimension, whereby the first dimension represents the regulatory standard support status of the vehicle product corresponding to the vehicle product demand analysis data in the target region. The second vehicle product decision factor is used to conduct a risk assessment of the vehicle product demand analysis data from a second dimension, whereby the second dimension represents the constraint standards related to the vehicle product. The third vehicle product decision factor is used to conduct a risk assessment of the vehicle product demand analysis data from a third dimension, whereby the third dimension represents the economic development status of the region where the vehicle product is used. The fourth vehicle product decision factor is used to conduct a risk assessment of the vehicle product demand analysis data from a fourth dimension, whereby the fourth dimension represents the cultural situation of the region where the vehicle product is used. The fifth vehicle product decision factor is used to conduct a risk assessment of the vehicle product demand analysis data from a fifth dimension, whereby the fifth dimension represents the technical standards and parameters required for the realization of the vehicle product. The sixth vehicle product decision factor is used to conduct a risk assessment of the vehicle product demand analysis data from a sixth dimension, whereby the sixth dimension represents the costs and benefits required for the realization of the vehicle product.

[0015] In one embodiment, the method further includes: constructing a large model system prompt; the large model system prompt includes a list of functions and available tools of the large model; the list of available tools includes a pre-built knowledge search tool, a network search tool, and a constraint and risk assessment tool; the functions include calling the knowledge search tool, calling the network search tool, and calling the constraint and risk assessment tool; wherein, the knowledge search tool is used to search from a pre-built vehicle knowledge base for first vehicle usage scenario information, first vehicle product project information, and first vehicle product constraint information associated with the vehicle product demand analysis data; the network search tool is used to search from the entire network for second vehicle usage scenario information, second vehicle product project information, and second vehicle product constraint information associated with the vehicle product demand analysis data; the constraint and risk assessment tool is used to determine the vehicle product risk assessment result corresponding to the vehicle product demand analysis data based on multiple vehicle product decision factors matching the vehicle product demand analysis data; and a large model for making vehicle product analysis decisions is constructed based on the large model system prompt.

[0016] In one embodiment, the method further includes: constructing a scenario library for recreating real-world demand scenarios through multi-dimensional vehicle usage scenarios, a project library for aligning demand analysis with actual goals, and a constraint library for verifying demand feasibility; performing data modeling and structured annotation on the original data associated with the scenario library, project library, and constraint library respectively, to obtain modeled structured data of the scenario library, modeled structured data of the project library, and modeled structured data of the constraint library; constructing an initial vehicle knowledge base based on the modeled structured data of the scenario library, modeled structured data of the project library, and modeled structured data of the constraint library; wherein the modeled structured data of the scenario library, modeled structured data of the project library, and modeled structured data of the constraint library in the vehicle knowledge base are associated based on storage identifiers; performing retrieval enhancement generation and parsing processing on the original data associated with the scenario library, project library, and constraint library respectively, to obtain the processed data corresponding to the scenario library, project library, and constraint library respectively, and uploading the processed data corresponding to the scenario library, project library, and constraint library respectively to the initial vehicle knowledge base to obtain the vehicle knowledge base.

[0017] In one embodiment, a pre-built knowledge search tool is invoked to search for first vehicle usage scenario information associated with vehicle product demand analysis data from a pre-built vehicle knowledge base. This includes: invoking the pre-built knowledge search tool to perform data augmentation processing on the vehicle product demand analysis data to obtain augmented vehicle product demand analysis data; determining a first set of candidate vehicle usage scenario information associated with the augmented vehicle product demand analysis data from the pre-built vehicle knowledge base; converting the augmented vehicle product demand analysis data into vector data; and determining a second set of candidate vehicle usage scenario information from the first set of candidate vehicle usage scenario information with a similarity greater than a preset similarity threshold to the vector data through a multi-path retrieval recall path; wherein different retrieval recall paths correspond to different recall methods; reordering multiple candidate vehicle usage scenario information in the second set of candidate vehicle usage scenario information; and determining the first vehicle usage scenario information associated with the vehicle product demand analysis data based on the reordered multiple candidate vehicle usage scenario information.

[0018] In one embodiment, a pre-built knowledge search tool is invoked to search for first vehicle product project information associated with vehicle product demand analysis data from a pre-built vehicle knowledge base. This includes: invoking the pre-built knowledge search tool to perform data augmentation processing on the vehicle product demand analysis data to obtain augmented vehicle product demand analysis data; determining a first set of candidate vehicle product project information associated with the augmented vehicle product demand analysis data from the pre-built vehicle knowledge base; converting the augmented vehicle product demand analysis data into vector data; and determining a second set of candidate vehicle product project information from the first set of candidate vehicle product project information with a similarity greater than a preset similarity threshold to the vector data through a multi-path retrieval recall path; wherein different retrieval recall paths correspond to different recall methods; reordering multiple candidate vehicle product project information in the second set of candidate vehicle product project information; and determining the first vehicle product project information associated with the vehicle product demand analysis data based on the reordered multiple candidate vehicle product project information.

[0019] In one embodiment, a pre-built knowledge search tool is invoked to search for first vehicle product constraint information associated with vehicle product demand analysis data from a pre-built vehicle knowledge base. This includes: invoking the pre-built knowledge search tool to perform data augmentation processing on the vehicle product demand analysis data to obtain augmented vehicle product demand analysis data; determining a first set of candidate vehicle product constraint information associated with the augmented vehicle product demand analysis data from the pre-built vehicle knowledge base; converting the augmented vehicle product demand analysis data into vector data; and determining a second set of candidate vehicle product constraint information from the first set of candidate vehicle product constraint information with a similarity greater than a preset similarity threshold to the vector data through a multi-path retrieval recall path; wherein different retrieval recall paths correspond to different recall methods; reordering multiple candidate vehicle product constraint information in the second set of candidate vehicle product constraint information; and determining the first vehicle product constraint information associated with the vehicle product demand analysis data based on the reordered multiple candidate vehicle product constraint information.

[0020] Secondly, embodiments of this application provide a vehicle product analysis and decision-making device, the device comprising:

[0021] The acquisition module is used to acquire the vehicle product demand analysis data input by the target object;

[0022] The processing module, based on a pre-built large model for vehicle product analysis and decision-making, invokes a pre-built knowledge search tool to search a pre-built vehicle knowledge base for first vehicle usage scenario information, first vehicle product project information, and first vehicle product constraint information related to vehicle product demand analysis data. The vehicle knowledge base includes vehicle usage scenario data for reconstructing real demand scenarios through multi-dimensional vehicle usage scenarios, vehicle product project data for aligning demand analysis with actual goals, and vehicle product constraint data for verifying demand feasibility. The first vehicle usage scenario information is obtained from the vehicle usage scenario data, the first vehicle product project information is obtained from the vehicle product project data, and the first vehicle product constraint information is obtained from the vehicle product constraint data.

[0023] The determination module, based on the first vehicle product constraint information, determines multiple vehicle product decision factors that match the vehicle product demand analysis data, and calls a pre-built constraint and risk assessment tool to determine the vehicle product risk assessment result corresponding to the vehicle product demand analysis data based on the multiple vehicle product decision factors.

[0024] The generation module is used to generate vehicle product analysis and decision results corresponding to vehicle product demand analysis data based on the first vehicle usage scenario information, the first vehicle product project information, and the vehicle product risk assessment results.

[0025] Thirdly, embodiments of this application provide an apparatus including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps of the method provided in the first aspect.

[0026] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method provided in the first aspect above.

[0027] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the method provided in the first aspect above.

[0028] The aforementioned vehicle product analysis and decision-making method, apparatus, and equipment include a computer device capable of acquiring vehicle product demand analysis data input by the target object; based on a pre-built large-scale model for vehicle product analysis and decision-making, it invokes a pre-built knowledge search tool to search a pre-built vehicle knowledge base for first vehicle usage scenario information, first vehicle product project information, and first vehicle product constraint information related to the vehicle product demand analysis data; the vehicle knowledge base includes vehicle usage scenario data for reconstructing real demand scenarios through multi-dimensional vehicle usage scenarios, vehicle product project data for aligning demand analysis with actual goals, and vehicle product constraint data for verifying demand feasibility; the first vehicle usage scenario data is used to retrieve vehicle product demand analysis data from a pre-built vehicle knowledge base. The scenario information is obtained from vehicle usage scenario data, the first vehicle product project information is obtained from vehicle product project data, and the first vehicle product constraint information is obtained from vehicle product constraint data. Based on the first vehicle product constraint information, multiple vehicle product decision factors that match the vehicle product demand analysis data are determined, and a pre-built constraint and risk assessment tool is invoked. Based on the multiple vehicle product decision factors, the vehicle product risk assessment result corresponding to the vehicle product demand analysis data is determined. Based on the first vehicle usage scenario information, the first vehicle product project information, and the vehicle product risk assessment result, the vehicle product analysis decision result corresponding to the vehicle product demand analysis data is generated. This method offers several advantages. First, the pre-built vehicle knowledge base includes vehicle usage scenario data for reconstructing real-world demand scenarios through multi-dimensional vehicle usage scenarios, vehicle product project data for aligning demand analysis with actual goals, and vehicle product constraint data for verifying demand feasibility. Therefore, by utilizing this pre-built knowledge base to identify and constrain vehicle product demand analysis data, not only can a basis for user demand analysis be provided, but policy and regulatory risks can also be mitigated, thereby improving the accuracy of subsequent vehicle product analysis and decision-making results. Second, by calling pre-built knowledge search tools and pre-built constraint and risk assessment tools for relevant processing, errors and illusions in the large model used for vehicle product analysis and decision-making during the analysis and summarization process can be reduced, thus improving the accuracy of vehicle product analysis and decision-making results. Attached Figure Description

[0029] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0030] Figure 1This is a flowchart illustrating a vehicle product analysis and decision-making method provided in an embodiment of this application;

[0031] Figure 2 This is a schematic diagram illustrating a process for constructing a vehicle product demand analysis and decision-making process based on a large model, as provided in an embodiment of this application.

[0032] Figure 3 This is a schematic diagram illustrating a data deduplication process provided in an embodiment of this application;

[0033] Figure 4 This is a schematic diagram of a data classification process provided in an embodiment of this application;

[0034] Figure 5 This is a schematic diagram of a data entry process provided in an embodiment of this application;

[0035] Figure 6 This is a schematic diagram illustrating the construction process of a vehicle knowledge base provided in an embodiment of this application;

[0036] Figure 7 This is a schematic diagram of a knowledge search process provided in an embodiment of this application;

[0037] Figure 8 This is a flowchart illustrating another vehicle product analysis and decision-making method provided in an embodiment of this application;

[0038] Figure 9 This is a schematic diagram of the structure of a vehicle product analysis and decision-making device provided in an embodiment of this application;

[0039] Figure 10 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0040] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0041] The vehicle product analysis and decision-making method provided in the embodiments of this application will be described below.

[0042] Please see Figure 1 , Figure 1 This is a flowchart illustrating a vehicle product analysis and decision-making method provided in an embodiment of this application. The method can be executed by a computer device. Figure 1 As shown, the vehicle product analysis and decision-making method may include, but is not limited to, the following steps:

[0043] S101. Obtain the vehicle product demand analysis data input by the target object.

[0044] In one alternative implementation, the vehicle product demand analysis data may include, but is not limited to, text data, image data, voice data, etc.

[0045] For example, vehicle product demand analysis data could be the text data such as "in-vehicle 220V inverter demand".

[0046] S102. Based on the pre-built large model for vehicle product analysis and decision-making, call the pre-built knowledge search tool to search the pre-built vehicle knowledge base for first vehicle usage scenario information, first vehicle product project information, and first vehicle product constraint information that are related to the vehicle product demand analysis data.

[0047] The vehicle knowledge base includes vehicle usage scenario data for reconstructing real-world demand scenarios through multi-dimensional vehicle usage scenarios, vehicle product project data for aligning demand analysis with actual goals, and vehicle product constraint data for verifying the feasibility of demands. The first vehicle usage scenario information is obtained by searching the vehicle usage scenario data, the first vehicle product project information is obtained by searching the vehicle product project data, and the first vehicle product constraint information is obtained by searching the vehicle product constraint data.

[0048] Specifically, vehicle usage scenario data, used to recreate real-world demand scenarios through multi-dimensional vehicle usage scenarios, can be stored in a scenario library; vehicle product project data, used to align demand analysis with actual goals, can be stored in a project library; and vehicle product constraint data, used to verify the feasibility of demands, can be stored in a constraint library. In other words, the vehicle knowledge base can include a scenario library, a project library, and a constraint library. The scenario library stores multiple vehicle usage scenario data sets, the project library stores multiple vehicle product project data sets, and the constraint library stores multiple vehicle product constraint data sets.

[0049] Optionally, vehicle usage scenario data can be constructed by computer equipment using historical user survey data, historical vehicle usage scenario data, historical decision-making data, and publicly available vertical website crawler data.

[0050] Optionally, vehicle product project data can be constructed using historical project implementation data such as historical project initiation, historical process specifications, historical project results, and historical summary reviews.

[0051] Optionally, vehicle product constraint data can be constructed by collecting existing industry constraint standards, regulatory standards, economic and cultural standards, and technical standards. Among these, constraint standards are used to determine the legality of vehicle products corresponding to the vehicle product demand analysis data; regulatory standards are used to characterize the management of vehicle products corresponding to the vehicle product demand analysis data in the target region.

[0052] Among them, the knowledge search tool can be the knowledge search interface corresponding to the vehicle knowledge base.

[0053] S103. Based on the first vehicle product constraint information, determine multiple vehicle product decision factors that match the vehicle product demand analysis data, and call the pre-built constraint and risk assessment tool to determine the vehicle product risk assessment result corresponding to the vehicle product demand analysis data based on the multiple vehicle product decision factors.

[0054] For example, suppose the first vehicle product constraint information includes "The automotive industry follows the ISO-26262 functional safety standard, and safety-related requirements have the highest priority in any vehicle model. This function must meet safety requirements" and "No relevant regulatory standards." In this case, the computer equipment can determine multiple vehicle product decision factors, including constraint standards and regulatory standards, that match the vehicle product requirements analysis data.

[0055] Among them, the constraint and risk assessment tool is used to constrain and assess the risks of vehicle product demand analysis data.

[0056] In one optional implementation, the computer device determines the vehicle product risk assessment result corresponding to the vehicle product demand analysis data based on multiple vehicle product decision factors. This can be achieved by: determining the vehicle product risk assessment score and weight corresponding to each of the multiple vehicle product decision factors; and performing a weighted summation of the vehicle product risk assessment scores corresponding to each of the multiple vehicle product decision factors to obtain the vehicle product risk assessment result corresponding to the vehicle product demand analysis data.

[0057] S104. Based on the first vehicle usage scenario information, the first vehicle product project information, and the vehicle product risk assessment results, generate the vehicle product analysis decision results corresponding to the vehicle product demand analysis data.

[0058] In one alternative implementation, the vehicle product analysis and decision-making results may include a summary of demand analysis, vehicle product risk assessment results, implementation recommendations, etc.

[0059] In this embodiment, the computer device can acquire vehicle product demand analysis data input by the target object; based on a pre-built large model for vehicle product analysis and decision-making, it calls a pre-built knowledge search tool to search for first vehicle usage scenario information, first vehicle product project information, and first vehicle product constraint information related to the vehicle product demand analysis data from a pre-built vehicle knowledge base; the vehicle knowledge base includes vehicle usage scenario data for reconstructing real demand scenarios through multi-dimensional vehicle usage scenarios, project data for aligning demand analysis and actual goals, and vehicle product constraint data for verifying demand feasibility; the first vehicle usage scenario information is obtained from the vehicle usage scenario data, the first vehicle product project information is obtained from the vehicle product project data, and the first vehicle product constraint information is obtained from the vehicle product constraint data; based on the first vehicle product constraint information, multiple vehicle product decision factors matching the vehicle product demand analysis data are determined, and a pre-built constraint and risk assessment tool is called to determine the vehicle product risk assessment result corresponding to the vehicle product demand analysis data based on the multiple vehicle product decision factors; based on the first vehicle usage scenario information, the first vehicle product project information, and the vehicle product risk assessment result, a vehicle product analysis decision result corresponding to the vehicle product demand analysis data is generated. This method offers several advantages. First, the pre-built vehicle knowledge base includes vehicle usage scenario data for reconstructing real-world demand scenarios through multi-dimensional vehicle usage scenarios, project data for aligning demand analysis with actual goals, and vehicle product constraint data for verifying demand feasibility. Therefore, by utilizing this pre-built vehicle knowledge base to identify and constrain vehicle product demand analysis data, not only can a basis for user demand analysis be provided, but policy and regulatory risks can also be mitigated, thereby improving the accuracy of subsequent vehicle product analysis and decision-making results. Second, by calling pre-built knowledge search tools and pre-built constraint and risk assessment tools for relevant processing, errors and illusions in the large model used for vehicle product analysis and decision-making during the analysis and summarization process can be reduced, thus improving the accuracy of vehicle product analysis and decision-making results.

[0060] In one alternative implementation, Figure 1In the vehicle product analysis and decision-making method shown, the computer device can also call a pre-built network search tool to search the entire network for second vehicle usage scenario information, second vehicle product project information, and second vehicle product constraint information related to the vehicle product demand analysis data. In this case, step S103, that is, the computer device determines multiple vehicle product decision factors that match the vehicle product demand analysis data based on the first vehicle product constraint information, can be: extracting multiple vehicle product decision factors that match the vehicle product demand analysis data from the first vehicle product constraint information and the second vehicle product constraint information; step S104, that is, the computer device generates the vehicle product analysis and decision-making result corresponding to the vehicle product demand analysis data based on the first vehicle usage scenario information, the first vehicle product project information, and the vehicle product risk assessment result, can be: generating the vehicle product analysis and decision-making result corresponding to the vehicle product demand analysis data based on the first vehicle usage scenario information, the first vehicle product project information, the second vehicle usage scenario information, the second vehicle product project information, and the vehicle product risk assessment result.

[0061] Among these, the online search tool can be a publicly available service provider interface.

[0062] Among them, the second vehicle usage scenario information, second vehicle product project information, and second vehicle product constraint information that are related to the vehicle product demand analysis data searched from the entire network can serve as supplementary information to the first vehicle usage scenario information, first vehicle product project information, and first vehicle product constraint information searched from the vehicle knowledge base.

[0063] It should be noted that the second vehicle usage scenario information, second vehicle product project information, and second vehicle product constraint information, which supplement the first vehicle usage scenario information, first vehicle product project information, and first vehicle product constraint information, mainly supplement the first vehicle usage scenario information, first vehicle product project information, and first vehicle product constraint information in terms of content. For example, assuming that the first vehicle product constraint information includes two items (e.g., information A and information B), and assuming that the second vehicle product constraint information has three items (e.g., information 1, information 2, and information 3), where information 1 is the same as information A, and information 2 is the same as information B, then information 3 can be used as supplementary information to the first vehicle product constraint information.

[0064] It should be noted that the second vehicle usage scenario information was obtained through an online search, or rather, through web crawling; while the first vehicle usage scenario information was obtained by searching for vehicle usage scenario data included in a vehicle knowledge base. This vehicle usage scenario data may include data obtained from web crawling. Therefore, there may be overlap between the first and second vehicle usage scenario information. However, the more frequently duplicated data appears, the higher its importance and the greater the attention it receives. Furthermore, since the data in the vehicle knowledge base is manually verified before being added to the database, the computer will prioritize using the first vehicle usage scenario information when making subsequent analyses and decisions based on it. Therefore, even if the first and second vehicle usage scenario information overlap, it will not affect subsequent analyses and decisions.

[0065] It should be noted that although calling the online search tool will retrieve multiple pieces of vehicle usage scenario information, multiple pieces of vehicle product project information, and multiple pieces of vehicle product constraint information related to the vehicle product demand analysis data from the entire network, each type of information has a priority. By default, the priorities of the multiple pieces of information corresponding to each type of information retrieved from the online search are accurate. Therefore, the second vehicle usage scenario information, the second vehicle product project information, and the second vehicle product constraint information determined by the computer device are respectively the Top-N vehicle usage scenario information, the Top-N vehicle product project information, and the Top-N vehicle product constraint information determined from the multiple pieces of vehicle usage scenario information, the multiple pieces of vehicle product project information, and the multiple pieces of vehicle product constraint information retrieved from the entire network.

[0066] It should be noted that if multiple vehicle product constraint information results in conflict among multiple results for a certain decision factor retrieved from the entire internet, the computer device will default to using the vehicle product constraint information corresponding to the decision factor most recently generated. For example, assuming the decision factor is regulatory standards, if constraint information 1 found in the internet search indicates that the regulatory standard supports the decision, and constraint information 1 was generated on November 6, 2025, while constraint information 2 indicates that the regulatory standard does not support the decision, and constraint information 2 was generated on October 6, 2024, the computer device may default to using constraint information 1. Furthermore, if a second vehicle product constraint information obtained from the internet search conflicts with a first vehicle product constraint information found in the knowledge base, the computer device will default to using the first vehicle product constraint information.

[0067] For example, suppose the first vehicle product constraint information includes "The automotive industry follows the ISO-26262 functional safety standard, safety-related requirements have the highest priority in any vehicle model, and this function must meet safety requirements" and "No relevant regulatory standards exist." Suppose the second vehicle product constraint information includes "This function has high acceptance" and "This function is a differentiating feature of the vehicle, with increased costs, and requires a certain level of economic status from the target buyers." In this case, the computer equipment can determine multiple vehicle product decision factors that match the vehicle product demand analysis data, including constraint standards, regulatory standards, culture, and economics. Among them, the vehicle product decision factor corresponding to "The automotive industry follows the ISO-26262 functional safety standard, safety-related requirements have the highest priority in any vehicle model, and this function must meet safety requirements" is constraint standards; the vehicle product decision factor corresponding to "No relevant regulatory standards exist" is regulatory standards; the vehicle product decision factor corresponding to "This function has high acceptance" is culture; and the vehicle product decision factor corresponding to "This function is a differentiating feature of the vehicle, with increased costs, and requires a certain level of economic status from the target buyers" is economics.

[0068] In this implementation method, the computer device can use a pre-built network search tool to search the entire network for second vehicle usage scenario information, second vehicle product project information, and second vehicle product constraint information that are related to the vehicle product demand analysis data. This can supplement the first vehicle usage scenario information, first vehicle product project information, and first vehicle product constraint information retrieved from the vehicle knowledge base, thereby improving the accuracy of the subsequent vehicle product risk assessment results and, consequently, the accuracy of the subsequent vehicle product analysis and decision-making results.

[0069] In one alternative implementation, Figure 1 In the vehicle product analysis and decision-making method shown, in step S103, the vehicle product risk assessment result may include the target vehicle product risk assessment score corresponding to the vehicle product demand analysis data. The computer equipment determines the vehicle product risk assessment result corresponding to the vehicle product demand analysis data based on multiple vehicle product decision factors in the following ways: for each vehicle product decision factor among the multiple vehicle product decision factors, obtain the target weight and vehicle product risk assessment score corresponding to the vehicle product decision factor; the target weight is preset or determined based on a pre-determined weight allocation strategy; based on the target weight and vehicle product risk assessment score corresponding to each vehicle product decision factor, determine the vehicle product risk assessment result corresponding to the vehicle product demand analysis data.

[0070] In some embodiments, the target weights corresponding to vehicle product decision factors can be determined by computer equipment based on vehicle product demand analysis data input by the target object. That is, the vehicle product demand analysis data input by the target object includes multiple vehicle product decision factors and the weights corresponding to each vehicle product decision factor.

[0071] For example, the target-corresponding input vehicle product demand analysis data could be: "Analyze the demand for 220V inverters in 30W models in China, where the target weights for the vehicle product decision factors constrained by regulatory standards, constraint standards, economic, cultural, technical standards, and cost-benefit factors are 0.2, 0.2, 0.1, 0.1, 0.2, and 0.2, respectively."

[0072] In some embodiments, the target weights corresponding to the vehicle product decision factors can also be default values. The default value is the average of the number of vehicle product decision factors currently involved in the calculation. For example, assuming there are four vehicle product decision factors currently involved in the calculation—regulatory standards, constraint standards, economic standards, and cultural standards—then the target weights corresponding to each of the four vehicle product decision factors would be 1 / 4 = 0.25.

[0073] In some embodiments, the target weights corresponding to vehicle product decision factors can also be determined by computer equipment in the following ways: based on the attributes corresponding to the vehicle product decision factors, determining the category to which the vehicle product decision factors belong, and determining the decision level to which the vehicle product decision factors belong; based on the decision level, determining the initial weights of the vehicle product decision factors; and based on the category to which the vehicle product decision factors belong and the vehicle product risk assessment score corresponding to the vehicle product decision factors, determining the weight correction coefficients corresponding to the vehicle product decision factors; and correcting the initial weights based on the weight correction coefficients to obtain the target weights corresponding to the vehicle product decision factors. In this way, the target weights of the vehicle product decision factors can be dynamically adjusted, providing an accurate data foundation for subsequently determining the vehicle product risk assessment results corresponding to the vehicle product demand analysis data, thereby improving the accuracy of the subsequently determined vehicle product risk assessment results.

[0074] For example, suppose the decision factors for vehicle products include regulatory standards, constraint standards, economic, cultural, and technical standards, and cost-benefit; where the attributes corresponding to regulatory standards and constraint standards are negative indicators, the attributes corresponding to economic, cultural, and technical standards are linearly correlated indicators, and the attributes corresponding to cost-benefit are non-linear gain indicators. In this case, the computer equipment can determine that the category to which the regulatory standards and constraint standards belong is the basic demand category, the category to which the economic, cultural, and technical standards belong is the expected demand category, and the category to which cost-benefit belongs is the exciting demand category.

[0075] Optionally, the computer equipment can determine the decision level corresponding to the vehicle product decision factors by using the Analytic Hierarchy Process (AHP). The AHP is a multi-criteria decision analysis method that combines qualitative and quantitative approaches. Its core idea is to decompose complex decision problems into levels such as objectives, criteria, and alternatives. By comparing each factor pairwise, the relative weight of each factor is calculated, thus providing a quantitative basis for selecting the optimal solution.

[0076] For example, assuming that vehicle product decision factors include regulatory standards, constraint standards, economic, cultural, technical standards, and cost-benefit, then computer equipment using the Analytic Hierarchy Process (AHP) can obtain the following hierarchical structure: [Target Layer: Functional Feasibility] --> B [Criterion Layer: Policy and Regulations] A --> C [Criterion Layer: Market Environment] A --> D [Criterion Layer: Technical Implementation] B --> B1 (Regulatory Standards) B --> B2 (Constraint Standards) C --> C1 (Economic) C --> C2 (Culture) D --> D1 (Technical Standards) D --> D2 (Cost-Benefit).

[0077] For example, following the example above, assuming that vehicle product decision factors include regulatory standards, constraint standards, economic, cultural, technical standards, and cost-benefit, the initial weights corresponding to each vehicle product decision factor can be shown in Table 1 below.

[0078] Table 1. Correspondence between decision factors and initial weights for each vehicle product

[0079]

[0080] Optionally, the computer equipment determines the weight correction coefficient corresponding to the vehicle product decision factor based on the category to which the vehicle product decision factor belongs and the vehicle product risk assessment score corresponding to the vehicle product decision factor. This can be achieved by obtaining the weight correction coefficient corresponding to the category to which the vehicle product decision factor belongs and the vehicle product risk assessment score corresponding to the vehicle product decision factor based on the correspondence relationship. The correspondence relationship includes the correspondence relationship between multiple combinations and multiple weight correction coefficients. Each combination includes the vehicle product risk assessment score range and the category to which the vehicle product decision factor belongs.

[0081] The correspondence can be a table pre-installed on a computer device (denoted as the correspondence table), or a table pre-installed in a database that the computer device can read (denoted as the correspondence table), etc., without limitation here. The correspondence table includes the correspondence between multiple combinations and multiple weight correction coefficients. Each combination includes the vehicle product risk assessment score range and the category to which the vehicle product decision factor belongs. For example, the correspondence table can be shown in Table 2 below.

[0082] Table 2 Correspondence Table

[0083]

[0084] Table 2 above can also be referred to as the Kano dynamic correction matrix.

[0085] Optionally, the computer equipment can correct the initial weights based on the weight correction coefficient to obtain the target weights corresponding to the vehicle product decision factors. This can be achieved by multiplying the initial weights and the weight correction coefficients as the target weights corresponding to the vehicle product decision factors.

[0086] For example, suppose the target object inputs the following vehicle product demand analysis data: "Analyze the demand for 220V inverters in 30W models in China", and suppose the vehicle product decision factors and the corresponding vehicle product risk assessment scores are as follows: (1) Regulatory standards: No relevant regulatory standards (6 points); (2) Constraint standards: Follow ISO-26262 functional safety standards (8 points); (3) Economy: Developed countries or regions (9 points); (4) Culture: High acceptance of new technologies (8 points); (5) Technical standards: Mature solutions are available and the difficulty is moderate (9 points); (6) Cost-benefit: ROI = 4.2 (explicit 4 points) + implicit benefit 3 points. Then the computer equipment can determine the vehicle product risk assessment results corresponding to the vehicle product demand analysis data based on the following pseudocode.

[0087] #Pseudocode Example

[0088] if national_policy_score == 0:

[0089] weights = [1.0, 0, 0, 0, 0, 0] # Regulatory standards rejected

[0090] else if law_regulation_score == 0:

[0091] weights = [0, 1.0, 0, 0, 0, 0] # Constraint criteria rejected

[0092] else:

[0093] weights = normalize([0.25, 0.25, 0.15, 0.1, 0.15, 0.1]) # Apply Kano correction

[0094] weights[0] *= 1.2 # Regulatory standard 6 points → Basic type 1.2 times

[0095] weights[5] *= 1.5 #Cost-benefit ratio 7 points → Excitement type 1.5 times

[0096] In this case, the computer equipment can determine that the target vehicle product risk assessment score (or feasibility score) included in the vehicle product risk assessment result corresponding to the vehicle product demand analysis data is 0.25×1.2×6+0.25×8+0.15×9+0.1×8+0.15×9+0.1×7×1.5=8.35; the risk assessment conclusion (or feasibility conclusion) included in the vehicle product risk assessment result is basically feasible, and implementation is recommended.

[0097] It should be noted that among the multiple vehicle product decision factors, the vehicle product risk assessment score corresponding to each factor is obtained by computer equipment through a comprehensive evaluation of multiple vehicle product constraint information related to the decision factor, within the first and second vehicle product constraint information. The comprehensive evaluation strategy can be determined based on expert experience or historical testing; no specific limitation is imposed here.

[0098] It should be noted that the sum of the target weights corresponding to multiple vehicle product decision factors does not necessarily have to be 1. The main purpose of adjusting the initial weights of each vehicle product decision factor is to make the final target vehicle product risk assessment score more realistic. For example, assuming the cost-benefit decision factor corresponds to a vehicle product risk assessment score of 7, this indicates that if the vehicle product is used in reality, it will receive positive feedback in terms of cost-benefit. However, since the initial weight of the cost-benefit decision factor is low, its influence on the target vehicle product risk assessment score can be increased by increasing its weight, thus making the target vehicle product risk assessment score more realistic.

[0099] Optionally, the computer device can also pre-configure the priority of the effective strategy for the target weights corresponding to the vehicle product decision factors; wherein, the weight effective strategy has the highest priority for user input, the next highest priority for scenario dynamic weight strategy, and the lowest priority for default strategy.

[0100] Optionally, users can define other vehicle product decision factors besides the multiple vehicle product decision factors, along with their corresponding target weights and vehicle product risk assessment scores, according to their scenario needs. The total score range is 0-10 points; users can simply input their requests. For example, a user might input, "Analyze the demand for 220V inverters in domestic 30W vehicle models. Please add a user pain point vehicle product decision factor with the weight: xxx, and its scoring criteria: xxx."

[0101] In some embodiments, the risk assessment results for vehicle products may further include a risk assessment conclusion, wherein the risk assessment conclusion is related to the risk assessment score of the target vehicle product. For example, if the risk assessment score of the target vehicle product is less than 4 points, the risk assessment conclusion is infeasible with a high risk; if the risk assessment score of the target vehicle product is 5-8 points, the risk assessment conclusion is basically feasible with a low risk; if the risk assessment score of the target vehicle product is 9 points or above, the risk assessment conclusion is feasible with no risk.

[0102] This implementation method can improve the accuracy of vehicle product risk assessment results corresponding to the determined vehicle product demand analysis data.

[0103] In one alternative implementation, Figure 1 In the vehicle product analysis and decision-making method shown, in step S103, multiple vehicle product decision factors may include at least two of the following: a first vehicle product decision factor, a second vehicle product decision factor, a third vehicle product decision factor, a fourth vehicle product decision factor, a fifth vehicle product decision factor, and a sixth vehicle product decision factor. The first vehicle product decision factor is used to conduct risk assessment on the vehicle product demand analysis data from a first dimension, whereby the first dimension represents the policy support situation for the vehicle product corresponding to the vehicle product demand analysis data in the target region. The second vehicle product decision factor is used to conduct risk assessment on the vehicle product demand analysis data from a second dimension, whereby the second dimension represents the constraint standards related to the vehicle product. The three vehicle product decision factors are: the third, the fourth, and the sixth. The third vehicle product decision factor is used to assess the risk of vehicle product demand analysis data from a third dimension, representing the economic development of the region where the vehicle product is used. The fourth vehicle product decision factor is used to assess the risk of vehicle product demand analysis data from a fourth dimension, representing the cultural situation of the region where the vehicle product is used. The fifth vehicle product decision factor is used to assess the risk of vehicle product demand analysis data from a fifth dimension, representing the technical standards and parameters required for the realization of the vehicle product. The sixth vehicle product decision factor is used to assess the risk of vehicle product demand analysis data from a sixth dimension, representing the costs and benefits required for the realization of the vehicle product.

[0104] Among them, the first vehicle product decision factor can be considered as the regulatory standard, and the corresponding standard includes: support, no support, no relevant standard - no risk, no relevant standard - risky; the corresponding vehicle product risk assessment score is 0-10 points.

[0105] Among them, the second vehicle product decision factor can be regarded as the constraint standard, and the corresponding standard includes: support, no support, no relevant constraint standard - no risk, no relevant constraint standard - risky; the corresponding vehicle product risk assessment score is 0-10 points.

[0106] Among them, the third vehicle product decision factor can be considered as the economy, which is used to characterize the economic development status and trend of the country or region that uses the vehicle product functions corresponding to the vehicle product demand analysis data, and corresponds to the total score of acceptance of payment; its corresponding vehicle product risk assessment score is 0-10 points.

[0107] Among them, the fourth vehicle product decision factor can be considered as culture, which represents the history, culture and ideology of the country or region where the vehicle product functions are used in accordance with the vehicle product demand analysis data, and the total score of the acceptance of the corresponding functions; its corresponding vehicle product risk assessment score is 0-10 points.

[0108] Among them, the fifth vehicle product decision factor can be considered as the technical standard, which represents the technical standards and parameters required to realize the vehicle product function corresponding to the vehicle product demand analysis data. The standard includes the difficulty of realizing the vehicle product function; its corresponding vehicle product risk assessment score is 0-10.

[0109] Among them, the sixth vehicle product decision factor can be considered as cost-benefit, which represents the cost and benefits required to realize the vehicle product functions corresponding to the vehicle product demand analysis data, and corresponds to the overall product competitiveness score; its corresponding vehicle product risk assessment score is 0-10 points.

[0110] It is understandable that the higher the risk assessment score of the vehicle product corresponding to the above vehicle product decision factors, the greater the corresponding risk or difficulty.

[0111] The computer equipment can determine the vehicle product risk assessment score corresponding to each of the above vehicle product decision factors based on the following predefined scoring rules:

[0112] (1) Regulatory Standards: The scoring criteria include: Support (8-10 points, automatically scored based on the strength of the specific support standard), No Support (0 points, automatically scored based on the specific opposition standard), No Relevant Standard - No Risk (4-7 points, automatically scored based on existing industry standards and regulatory data), No Relevant Standard - Risky (1-3 points, automatically scored based on existing industry standards and regulatory data). If no relevant standard is found, the default is No Relevant Standard - No Risk, with a default score of 6 points;

[0113] (2) Constraint Criteria: The scoring criteria include: Support (8-10 points, automatically scored based on the strength of support for the specific constraint criteria), No Support (0 points, automatically scored based on the specific constraint criteria against), No Relevant Constraint Criteria - No Risk (4-7 points, automatically scored based on existing industry standards), and No Relevant Constraint Criteria - Risky (1-3 points, automatically scored based on existing industry standards). If no relevant constraint criteria are found, the default is no relevant constraint criteria - no risk, with a default score of 6 points.

[0114] (3) Economy: The score is assessed based on the economic development trend of the country and region, including developed regions (9-10 points), middle-income regions (6-8 points), underdeveloped regions (3-5 points), and poor regions (0-2 points). If the local economic level is on an upward or downward trend, the score will be increased or decreased by 1-2 points based on the existing score. The evaluation criteria are mainly based on the standards formulated by relevant departments or publicly available data. If there is no relevant evaluation data, the default score is 6 points.

[0115] (4) Culture: The evaluation will be based on the local history, culture, and ideology. Any conflict with these aspects will receive 0 points. A positive attitude towards new functional technologies without any conflict with humanistic ideology will receive 6-10 points. A negative attitude towards new functional technologies or any conflict with humanistic ideology will receive 1-5 points. Specific evaluations will be based on the survey results. If no relevant evaluation data is available, the default score will be 5 points.

[0116] (5) Technical Standards: The score is based on the maturity and difficulty of the technical solution for implementing the function. A score of 0 is given if the function cannot be implemented. The scoring criteria include: New technology - low implementation difficulty (8-10 points), New technology - high implementation difficulty (5-7 points), Old technology - low implementation difficulty (4-6 points), Old technology - high implementation difficulty (1-3 points). For functional scenarios where the technical standards cannot be evaluated in the early stages, the technical standards can be excluded and not evaluated for the time being.

[0117] (6) Cost and Benefit: A comprehensive score comparing the costs and benefits of technology implementation (including explicit benefits such as revenue and implicit benefits such as user reviews, technological leadership, and company image). The score includes explicit benefits (6 points) and implicit benefits (4 points), which are scored separately and then added together.

[0118] Explicit benefit calculation: Benefit ratio = Benefit / Cost. Benefit ratio range: greater than 5 gets 6 points, 1-5 gets 1-5 points, less than 1 gets 0 points. For functional scenarios where benefits cannot be assessed in the early stages, explicit benefits can be excluded and not evaluated for the time being.

[0119] Implicit revenue calculation: Implicit revenue is mainly determined by user ratings of features and technology. All positive reviews earn 4 points, over two-thirds positive reviews earn 3 points, over half positive reviews earn 2 points, over one-third positive reviews earn 1 point, and less than one-third positive reviews earn 0 points. For features whose revenue cannot be assessed initially, implicit revenue can be excluded and not evaluated.

[0120] It should be noted that if regulatory or constraint standards are not supported, no points will be awarded, and the function cannot be launched in the current region. Risk issues will be assessed from 1 to 7 points based on the severity of the risk. No regulatory or constraint standards will result in a default score of 6 points. If regulatory and constraint standards are supported, the score will be 8-10 points based on the level of support.

[0121] By adopting this implementation method, computer equipment can conduct risk assessments on vehicle products or vehicle product functions corresponding to vehicle product demand analysis data from multiple dimensions, thereby improving the accuracy of the corresponding vehicle product risk assessment results.

[0122] In one alternative implementation, Figure 1In the vehicle product analysis and decision-making method shown, the computer equipment can also construct large model system prompts. These prompts include a list of large model functions and available tools. The list of available tools includes pre-built knowledge search tools, network search tools, and constraint and risk assessment tools. Functions include calling the knowledge search tools, the network search tools, and the constraint and risk assessment tools. Specifically, the knowledge search tools are used to search from a pre-built vehicle knowledge base for first vehicle usage scenario information, first vehicle product project information, and first vehicle product constraint information associated with the vehicle product demand analysis data. The network search tools are used to search from the entire network for second vehicle usage scenario information, second vehicle product project information, and second vehicle product constraint information associated with the vehicle product demand analysis data. The constraint and risk assessment tools are used to determine the vehicle product risk assessment result corresponding to the vehicle product demand analysis data based on multiple vehicle product decision factors matching the vehicle product demand analysis data. Based on the large model system prompts, a large model for vehicle product analysis and decision-making is constructed.

[0123] Optionally, the function name of the pre-built knowledge search tool can be: `search_knowledge_base`. Parameters can include: `query` (a string) and `request_type` (an optional string), where `query` represents the search keyword, such as "smart home user scenarios," and `request_type` specifies the search library type, which can include `scene` (scenario library), `project` (project library), and `constraint` (constraint library); if not specified, the entire library is searched by default. Its function is to retrieve matching vehicle usage scenario data (corresponding to the aforementioned first vehicle usage scenario information), historical project cases (corresponding to the aforementioned first vehicle product project information), or constraints (corresponding to the aforementioned first vehicle product constraint information) from the vehicle knowledge base, and return structured information (such as user requirements, functional cases, and regulatory clauses).

[0124] Optionally, the function name of the pre-built online search tool can be: search_web. Parameters can include: query (string), representing a web-wide search query, such as "latest data privacy standards". Its function is to supplement real-time external information (such as policy updates and industry trends) to fill the timeliness gaps in the vehicle knowledge base.

[0125] Optionally, the function name for the pre-built constraint and risk assessment tool can be: assessment_feasibility.

[0126] Parameters (all dictionary type, must include score, weight, and description) may include: national_policy: support for regulatory standards (0-10 points), law_regulation: compliance with constraint standards (0-10 points), economic: economic environment suitability (0-10 points), culture: cultural acceptance (0-10 points), technology_standard: difficulty of technical implementation (0-10 points), cost_benefit: cost-benefit balance (0-10 points). Its function is: based on the dynamic weighting rules of vehicle product decision factors, with all factor weights summed to 1, calculate the feasibility score (0-10 points) and output a risk conclusion (infeasible (4 points and below) / basically feasible (5-8 points) / feasible (9 points and above), where national_policy=0 or law_regulation=0 or culture=0 or technology_standard=0 directly results in 0 points, indicating infeasibility).

[0127] In some embodiments, the prompts in the large model system may also include role definition, workflow standardization, output requirements, exception handling mechanism, example interaction, etc.

[0128] For example, the role definition could be: You are a professional product analysis and decision engineer who focuses on mining user needs and analyzing scenario feasibility through multi-source data. Based on scenario knowledge bases and external information, you can help users conduct product requirement analysis and decision evaluation, and output structured decision recommendations.

[0129] For example, the workflow specification is as follows: When a user submits a requirement analysis request, please strictly follow these steps to call the tool and generate a response: Step 1, Requirement Analysis: Deconstruct the user query and extract key elements (such as product domain, target user, scenario type); Step 2, Data Collection: Prioritize calling search_knowledge_base to obtain historical data from the scenario library, project library, and constraint library. If the latest information is needed, call search_web to supplement the entire network data; Step 3, Factor Mapping: Extract vehicle product decision factors (such as policy terms and technical costs) from the search results (such as the decision basis found in the search) and format them into the parameters required by assess_feasibility; Step 4, Feasibility Assessment: Call assess_feasibility and output the score and risk level according to the rule chain; Step 5, Decision Generation: Analyze and integrate the collected information, combine the results returned by the tool, and give a conclusion (including requirement matching degree, constraint limitations, and improvement suggestions).

[0130] For example, the output requirements are as follows: 1. Decision recommendations should include: a summary of the needs analysis, feasibility assessment results, implementation suggestions, and risk warnings; 2. The content should be organized using Markdown format; 3. Key data should be presented in tables; 4. Risk assessments should use a scoring system (0-10 points).

[0131] For example, exception handling mechanisms include: 1. Terminating the generation operation and triggering a manual confirmation mechanism or retry when the toolchain call times out or fails; 2. Terminating the generation operation and triggering a manual confirmation mechanism or retry when the toolchain call parameters do not meet the tool's expectations; 3. Terminating the generation operation and triggering a manual confirmation mechanism or retry when the model does not execute according to the predetermined workflow specifications; 4. Terminating the generation operation and triggering a manual confirmation mechanism or retry when the workflow result is not returned as expected.

[0132] For example, a sample call might be: User query: "Analyze the demand for 220V in-vehicle inverters in domestic 300,000 RMB car models"; Model response flow:

[0133] 1. Call search_knowledge_base("In-vehicle 220V inverter power supply demand scenario", "scene") to obtain user demand and usage frequency data in similar scenarios.

[0134] 2. Call search_knowledge_base("In-vehicle 220V Inverter Power Supply Project", "project") to obtain existing project examples.

[0135] 3. Call search_knowledge_base("regulatory standards and constraints for 220V inverter power supply in vehicles", "constraint") to retrieve compliance requirements such as GDPR.

[0136] 4. Call search_web("Latest regulatory standards and constraint standards for 220V inverter power supply in vehicles") to update the constraint information dynamically.

[0137] 5. Factor extraction based on data:

[0138] national_policy: {score: 6, weight:0.3, description: "No relevant regulatory standards or restrictions"}

[0139] law_regulation: {score: 8, weight:0.2, description: "Must comply with ISO-26262 functional safety standard"}

[0140] ... (other factors)

[0141] Note: The weight value can be directly entered by the user during renewal. The default weight is the average.

[0142] 6. Calling assess_feasibility(...) yields a score of 7, with the conclusion "basically feasible".

[0143] 7. Based on the comprehensive analysis of data from the scenario library, project library, and constraint library, the final output is: "xxx".

[0144] Please see Figure 2 , Figure 2 This is a schematic diagram illustrating a process for constructing a vehicle product demand analysis and decision-making process based on a large model, as provided in an embodiment of this application. Figure 2As shown, (1) the computer device can obtain the dialogue input by the user on the dialogue webpage (i.e., vehicle product demand analysis data); (2) based on the dialogue input by the user, determine the user's question; (3) input the user's question into the large model; based on the large model structure, the user query extraction tool calls the parameter calling tool, including (4) calling the knowledge search tool to query the vehicle knowledge base for vehicle usage scenario data (corresponding to the first vehicle usage scenario information), project case data (corresponding to the first vehicle product project information) and vehicle product constraint data (corresponding to the first vehicle product constraint information) that match the user's question; (5) calling the network search tool to search the network for the second vehicle usage scenario information and the second vehicle (6) Extract constraint tool parameter results based on query results (including first vehicle usage scenario information, first vehicle product project information, first vehicle product constraint information, second vehicle usage scenario information, second vehicle product project information and second vehicle product constraint information), construct new prompt words, and reconstruct prompt words based on new prompt words; (7) Call constraint and risk assessment tools to determine the vehicle product or vehicle product function risk assessment results corresponding to the user problem from the dimensions of regulatory standards, constraint standards, technical standards, etc.; (8) Use the large model to conduct conclusion analysis based on all the results obtained above.

[0145] In some embodiments, the computer device may also pre-build a workflow specification for the large model. For example, the workflow specification for the large model may be as follows:

[0146] User input:

[0147] Analysis of the demand for 220V in-vehicle inverters in domestic 30W car models

[0148] Model output:

[0149] Based on the search results, I will analyze the current situation, influencing factors, and prospects of the demand for in-vehicle 220V inverters in domestic vehicles priced around 300,000 yuan. The analysis process references the structured methodology of keyword engineering, combining a scenario library (user usage scenarios), a project library (historical cases), and a constraint library (regulatory, technical, and economic limitations) for multi-dimensional evaluation. The following are the analysis conclusions:

[0150] ##Analysis of Background and Objectives

[0151] Context: The vehicle's 220V power supply converts the vehicle's 12V DC power to AC power via an inverter, supporting low-power appliances (such as car refrigerators, water heaters, and laptops), but with a power limit of approximately 150W. High-power devices (such as induction cookers) are not suitable. The current market for new energy vehicles in the 300,000 RMB price range is experiencing rapid growth, with users focusing on technological experiences and personalized needs.

[0152] Analysis Objective: To assess the demand intensity, feasibility, and implementation challenges of this feature across 300,000 vehicle models, providing a basis for product decision-making.

[0153] ## Demand Analysis: Explicit and Implicit Factors

[0154] 1. User scenarios drive demand

[0155] Explicit needs: In scenarios such as outdoor camping and business travel, there is a clear need for temporary power, such as connecting to a car refrigerator or charging electronic devices. Search results show that users are concerned about inverter safety (such as the risk of power overload) and cost (price range of 200-800 yuan), indicating that practical functionality is key.

[0156] Implicit demand: Car owners seek a "multi-functional" technological label, and a 220V power supply can enhance product differentiation and increase pricing power. For example, a certain brand's car model meets this demand through outdoor power accessories, indirectly strengthening the brand's technological image.

[0157] 2. Market Positioning and User Profile

[0158] The 300,000 model's users are primarily first- and second-tier city residents upgrading or replacing their vehicles, emphasizing intelligent features and scalability. However, their demand is stratified:

[0159] High-intensity groups: outdoor enthusiasts and business users, who consider it an essential need (such as boiling water during long-distance travel).

[0160] Low-intensity group: Daily commuters, who use the device infrequently, may consider it "icing on the cake".

[0161] 3. Historical Project Reference

[0162] Local case study: The "Model A vehicle equipped with an in-vehicle 220V inverter" project has been successfully implemented, and user feedback has been positive. It is recommended that this feature be rolled out to other vehicles in the future.

[0163] Positive examples found in online searches: The xx car power bank (1200W output) validated the market's acceptance of portable 220V solutions, but it requires external devices and is not integrated into the car.

[0164] Negative examples found online: XX car series removed the 220V socket from the car, mainly due to battery technology adjustments, cost control, and low-frequency usage data. This suggests that if the demand is not universal, integration may be counterproductive.

[0165] Feasibility Assessment: Constraints and Risks

[0166] Based on the constraint library model, key vehicle product decision factors are scored (0-10 points, with higher scores indicating lower risk):

[0167] Regulatory standards (score 6): No relevant regulatory standards.

[0168] Constraint Standards (Score 8): Comply with ISO-26262 functional safety standards and electrical safety standards (such as power limits for onboard electrical equipment).

[0169] Economy (score 9): It belongs to a developed country or region.

[0170] Culture (score 8): The trend of upgrading new energy consumption supports functional innovation and has a high acceptance of new technologies.

[0171] Technical Standards (Score 9): Inverter technology is mature, but in-vehicle integration requires consideration of circuit safety and heat dissipation design, which increases costs by about 500-1000 yuan per vehicle.

[0172] Cost-benefit analysis (score 7): The cost of adding new features may be passed on to the selling price, but users of vehicles priced at 300,000 RMB have low price sensitivity; the benefit is reflected in differentiated competition, but the frequency of use needs to be balanced (e.g., low-frequency scenarios may lower ROI). Overall score: ROI = 4.2 (explicit benefit 4 points) + implicit benefit 3 points.

[0173] Feasibility conclusion: The overall score is about 7.9 points (the weighting adopts a dynamic weighting strategy that combines the Kano model and the analytic hierarchy process), which is "basically feasible with low risk", but it is necessary to prioritize matching specific vehicle positioning (such as SUVs that focus on outdoor functions).

[0174] ## Conclusions and Recommendations

[0175] Overall demand is moderate: 220V inverter functionality has a niche demand in vehicles priced around 300,000 RMB, but it's not a universal selling point. It's more suitable for outdoor-oriented or commercial vehicles, increasing scenario coverage; for urban commuter vehicles, it's recommended to offer it as an optional feature or with an external connection to reduce redundancy costs.

[0176] Implementation strategy:

[0177] In the short term: Quantify the frequency of demand through user surveys (such as referring to in-vehicle electrical appliance usage data) and prioritize integration in high-end SUV versions.

[0178] Long-term: Combining advancements in battery technology (such as high discharge capacity) with exploration of intelligent power management to balance cost and user experience.

[0179] Risk warning: If the user scenario is not accurately identified, this feature may be canceled, resulting in a waste of resources.

[0180] Using this implementation method, a large model can be constructed for vehicle product analysis and decision-making, thereby preparing for subsequent vehicle product demand analysis and decision-making.

[0181] In one alternative implementation, Figure 1In the vehicle product analysis and decision-making method shown, the vehicle knowledge base mentioned in step S102 can be constructed by computer equipment in the following ways: Constructing a scenario library for recreating real demand scenarios through multi-dimensional vehicle usage scenarios, a project library for aligning demand analysis with actual goals, and a constraint library for verifying demand feasibility; performing data modeling and structured annotation on the original data associated with the scenario library, project library, and constraint library respectively, to obtain modeled structured data of the scenario library, modeled structured data of the project library, and modeled structured data of the constraint library; constructing an initial vehicle knowledge base based on the modeled structured data of the scenario library, modeled structured data of the project library, and modeled structured data of the constraint library; wherein, the modeled structured data of the scenario library, modeled structured data of the project library, and modeled structured data of the constraint library in the vehicle knowledge base are associated based on storage identifiers; performing retrieval enhancement generation and parsing processing on the original data associated with the scenario library, project library, and constraint library respectively, to obtain the processed data corresponding to the scenario library, project library, and constraint library respectively, and uploading the processed data corresponding to the scenario library, project library, and constraint library respectively to the initial vehicle knowledge base to obtain the vehicle knowledge base.

[0182] In some embodiments, computer devices may utilize historical user survey data, historical vehicle usage scenario data, historical decision-making data, and publicly available vertical website crawler data to construct a scenario library for reconstructing real-world demand scenarios through multi-dimensional vehicle usage scenarios. The characteristic information of the vehicle usage scenario data in the scenario library may include requirements, users, scenarios, functions, problem-solving, experiences, and decision suggestions.

[0183] For example, assuming a computer device builds a scenario library based on historical decision data, taking the "in-vehicle 220V inverter demand" analysis and decision-making process to construct feature data as an example, its JSON example is as follows:

[0184] {

[0185] "require": "The vehicle supports 220V power supply",

[0186] "requireSource": "Market Feedback",

[0187] "user": "Vehicle owner, passenger",

[0188] "scene": ["Car Owner Camping","Work Office - Commute","Work Office - Client Visit"],

[0189] "Industry": "Featured in high-end models, using multiple mature solutions, with costs reduced to below xx yuan."

[0190] "function": "In-vehicle 220V inverter power supply",

[0191] "solve": "Use of high-power electrical appliances in vehicles",

[0192] "case": "rice cooker, hair dryer, laptop",

[0193] "experience": {

[0194] "score": "7.5",

[0195] "frequency": "1-3 times / week",

[0196] "level": "high",

[0197] "cost": "Buyout (less than xx yuan)"

[0198] },

[0199] "attachment": "file"

[0200] "dataSource": "After-sales feedback"

[0201] }

[0202] In the JSON example above, the meanings of each field are explained as follows: Requirement: User-required feature; Requirement Source: Collection of user requirements; User: The target user group and user profile for this feature; Scene: The usage scenario category for this feature, subcategories can be added; Industry: Market conditions for this product feature, such as implementation cost, whether it is a first-of-its-kind feature, etc.; Function: The product function implemented by this requirement; Solve: The requirement problem solved by this product feature; Case: The problem solved by this product feature. Case Study; Experience: User experience of the product feature (does it solve user problems, and how is it rated); User Score: User rating (0-10); Frequency of Use: times / day, week, month, year; Level of Need: High, medium, low (users' level of need for the feature); Willingness to Pay: Per use (price range), outright purchase (price range), monthly subscription (price range); Attachments: Documents, images, audio and video, etc., related to the feature requirement research; Data Source: The source of the requirement, such as user feedback, website crawling, internal proposal, etc.

[0203] In some embodiments, computer equipment can utilize historical project implementation data, such as historical project initiation, historical process specifications, historical project results, and historical summary reviews, to construct a project library for aligning requirements analysis with actual goals. The characteristic information of vehicle product project data in the project library may include: project background, requirements analysis, competitiveness analysis, project target, technology solution, project period, and cost-benefit information. Taking "in-vehicle 220V inverter requirements" as an example, the JSON example of the project characteristic data is as follows:

[0204] {

[0205] "projectName": "Project requiring in-vehicle 220V inverter for vehicle model A",

[0206] "vehicleModel": "Vehicle Model A",

[0207] "priceRange": "450,000-600,000",

[0208] "background": "Market feedback indicates that customers have needs for power supply in vehicles and for parking / camping."

[0209] "requireAnalysis": "Car owners have a strong demand for external devices in camping / office scenarios, and there is a significant need for power and charging equipment while driving. Power needs also arise during parking in rain or snow, and short trips. Therefore, a 220V inverter power supply is required inside the vehicle to meet user needs."

[0210] "competitionity Analysis": "Competitors in the same vehicle models all feature this function. It can meet various in-vehicle power needs, improving user satisfaction. Several mature solutions are already available, with costs reduced to below xx yuan."

[0211] "target": "Develop and mass-produce a 220V in-vehicle inverter power supply function, functional parameters: xx",

[0212] "Technology": "1. Seven-in-one integrated electric drive. Advantages: xxx, Disadvantages: xxx; 2. External inverter with controller. Advantages: xxx, Disadvantages: xxx. Conclusion: Choose option 2, Reason: xxx, Risk solution: xxx",

[0213] "costFeasibility": "This solution, costing xx, meets the project requirements."

[0214] "period": "The project development cycle is xx months, from 2025.x to 2025.x".

[0215] "costBenefit": "Implementation of this technical solution: R&D cost: xxx; cost per vehicle: xxx; after-sales cost: xxx; total cost: xxx. Benefits: explicit benefits: xxx, implicit benefits: xxx."

[0216] "result": "Project successfully implemented, and has been installed on vehicle model A."

[0217] "reasonReview": "Requirements: Meets the needs of high-end users. Technology: Mature solution, simple implementation. Cost: Cost meets xxx. Regulations: xxx",

[0218] "attachment": "file"

[0219] }

[0220] In the JSON example above, the meanings of each field are explained as follows: Project Name: Project name; Vehicle Model: Vehicle model for which this product feature is supported; Price Range: Price range for the vehicle model for which this product feature is supported; Background: The background of the project, generally derived from user needs, business departments, company plans, etc.; Requirement Analysis: Evaluating the requirements proposed by users and business departments, confirming the connection between requirements and implementation solutions; Competitiveness Analysis: A competitive analysis of this product feature or technology against similar products. Strengths or weaknesses; Project target: The product functional goals that the project needs to achieve; Technology: The technical solutions, risks, and costs for implementing the function; Project period: The development time of the project; Cost-benefit: The cost and benefit value of implementing the function in the project, and its impact on the product; Result: Whether the project succeeded or failed; Reason review: Review of the reasons for the success or failure of the project, including requirements, technology, costs, policies, regulations, etc.; Attachment: Documents and images such as project initiation materials, prototype design, solution design, and product specifications.

[0221] In some embodiments, computer equipment can construct a constraint library for verifying the feasibility of requirements by collecting vehicle product constraint data such as existing industry constraint standards, national and regional policies, economic and cultural factors, and technical standards. The characteristic information of the vehicle product constraint data in the constraint library mainly includes information such as national policy, law regulation, economic, cultural, and technology standards.

[0222] For example, a computer device can construct vehicle product constraint data by organizing documents and corresponding metadata, as shown in the following JSON example:

[0223] {

[0224] "nationalPolicy": "This feature has no specific regulatory standards but must meet automotive-grade safety requirements. Evaluation score: 6 points."

[0225] "lawRegulation": "The automotive industry follows the ISO-26262 functional safety standard, and safety-related requirements have the highest priority in any vehicle model. This function must meet safety requirements, and the assessment score is 8 points."

[0226] "culture": "This feature is highly accepted. Rating: 8 out of 10."

[0227] "economic": "This feature is a high-end, differentiated feature of the vehicle, which increases costs and requires a certain economic level from the buyers. As an economically developed region, it scores 9 points."

[0228] "technologyStandard": "Complies with ISO-26262 functional safety standard and Q / SK J07.21-2020 electromagnetic compatibility requirements for electrical / electronic components and subsystems, mature solution architecture, and achieves a score of 9 out of 10 in difficulty assessment."

[0229] "costBenefit": "Implementation of this technical solution: R&D cost: xxx; cost per vehicle: xxx; after-sales cost: xxx; total cost: xxx. Benefits: explicit benefits: xxx, implicit benefits: xxx. Overall evaluation score: 7 points."

[0230] Conclusion: The overall score for this feature is xxx. Feasibility: Feasible; implementation is recommended.

[0231] }

[0232] In the JSON example above, the meanings of each field are explained as follows: Category: The category of this constraint, including national policy, law regulation, economic, culture, technology standard, etc.; Source: The source of this constraint, including policy regulations, technical specifications, and other documents; Validity Period: The validity period of this constraint; Attachment: The original source document of this constraint.

[0233] In some embodiments, the computer device performs data modeling and structured annotation on the original data associated with the scene library, project library, and constraint library respectively, to obtain the modeled structured data of the scene library, the modeled structured data of the project library, and the modeled structured data of the constraint library. The process of performing data modeling and structured annotation on the original data associated with each library may include, but is not limited to, the following steps:

[0234] Step 1: Perform data cleaning on the raw data to obtain cleaned data. This corrects or removes erroneous, incomplete, duplicate, or irrelevant parts, thereby improving the quality and usability of the data.

[0235] Optionally, the computer equipment may perform data cleaning on the raw data, including: using regular expressions to remove invisible characters and replace Unicode spaces with standard spaces \u0020; using a Traditional Chinese to Simplified Chinese service to replace Traditional Chinese with Simplified Chinese; removing Hyper Text Markup Language (HTML) tags, removing web page identifiers, and retaining only plain text content; using regular expressions to remove emoticons and remove privacy information (replacing email addresses, ID card numbers, bank card numbers, etc.); using regular expressions to match Uniform Resource Locators (URLs) and replace them with [URL] placeholders; and customizing other cleaning operations as needed, such as other removal or replacement operations.

[0236] Step 2: Perform deduplication on the cleaned data to obtain the deduplicated data. This improves data processing efficiency and accuracy.

[0237] Optionally, the cleaned data is text data; the process of deduplication of the cleaned data by computer equipment can be as follows: Figure 3 As shown, Figure 3 This is a schematic diagram illustrating a data deduplication process provided in an embodiment of this application. For example... Figure 3 As shown, the computer device can first perform word segmentation on the text data to obtain segmented text data; then, it can remove stop words from the segmented text data to obtain processed text data; determine the similarity between each pair of processed text data; determine a preset similarity threshold; and based on the preset similarity threshold and the similarity between the text data, remove highly similar text to obtain deduplicated data. The preset similarity threshold can be estimated by selecting a small amount of data for similarity calculation and based on the actual data situation.

[0238] For example, suppose the cleaned data includes Text 1:

XX Model

XX Model

[0239] Text 1: [XX Model] Luxurious six-seat spacious interior for a more comfortable driving experience!

[0240] Text 2: [XX Model] xx Intelligent Vehicle System, Enjoy a New Experience of Technological Driving!

[0241] Secondly, the computer device can perform stop word removal on the segmented text data to obtain the processed text data. Assuming the stop words are: "de", "le", "zai", punctuation, etc., the processed text data is as follows:

[0242] Text 1: XX model, luxurious, six-seater, extra-large, space, driving, travel, more comfortable

[0243] Text 2: XX model, xx intelligent, car machine system, enjoy technology, new driving experience

[0244] Then, the computer device can calculate the similarity between the processed Text 1 and the processed Text 2. Among them, the similarity between Text 1 and Text 2 (denoted as J(1,2)) can be calculated by the following formula.

[0245]

[0246] Through the above calculation, the computer device can determine that the similarity between Text 1 and Text 2 is 2 / 18 ≈ 0.12.

[0247] After that, assuming the preset similarity threshold is 0.5, the computer device can determine that the similarity 0.12 between Text 1 and Text 2 is less than the preset similarity threshold 0.5. In this case, the computer device can determine that Text 1 and Text 2 are not duplicate data and retain Text 1 and Text 2.

[0248] Step 3: Classify the data.

[0249] Optionally, the computer device can construct classification labels based on the data characteristics of the data, and can be specifically classified as follows:

[0250] Vehicle usage scenario data label: Classify by scenario, such as: going out - camping, going out - traveling, working - commuting to work, working - visiting customers, etc.;

[0251] Project data label: Classify by function, such as: software - cockpit, software - voice, hardware - vehicle power supply, hardware - intelligent chassis, etc.;

[0252] Vehicle product constraint data label: Classify by constraint, such as: supervision, regulations, culture, etc.

[0253] Optionally, when the computer device classifies the deduplicated data, it can compare the similarity between the data meaning (such as vehicle usage scenario data) or data metadata (such as vehicle product constraint data) and the label data, and classify based on the comparison result. This process can be as Figure 4 shown, Figure 4 is a schematic diagram of a data classification process provided by an embodiment of the present application. As Figure 4As shown, the computer device can collect corresponding domain data based on data tags, perform word segmentation on the collected corresponding tag domain data to obtain segmented data; remove stop words from the segmented data to obtain deduplicated data; construct a keyword thesaurus and frequency information table; determine the keyword frequency threshold based on the keyword thesaurus and frequency information table; determine the cosine similarity between the deduplicated data and the keywords; and select the word with the highest similarity as the classification label for the corresponding tag domain data.

[0254] For example, the process by which computer devices classify data is as follows:

[0255] Tag-based data:

[0256]

[0257] Word segmentation and stop word removal:

[0258]

[0259] Build a keyword thesaurus and frequency table:

[0260]

[0261] The computer device can set a keyword frequency threshold (e.g., greater than 1) based on a word frequency table to select keywords as the keyword list for this tag. Optionally, the computer device can also adjust the word frequency threshold according to actual needs.

[0262] Next, the computer can calculate the similarity based on the deduplicated word text, taking text 1 as an example. The cosine similarity of text 1 and the vocabulary list labeled "outing-camping" is calculated, yielding a similarity value of 0.316. This method can be used to calculate the similarity of all data with their corresponding labels, ultimately determining the classification of the data. If the similarity of all calculated labels for a given data point is lower than a preset minimum similarity threshold, the data is considered not to be in the current label library and must be classified into a default label, such as "other". The minimum threshold is an empirical value and can be referenced from text similarity threshold setting schemes.

[0263] Optionally, the computer equipment can re-classify and calculate the entire dataset after new label data is added to ensure data accuracy. Optionally, the computer equipment can use a scheduled batch update strategy to update the label data.

[0264] Step 4: Data labeling.

[0265] Optionally, the computer equipment can perform data modeling and structured annotation on the processed data according to the corresponding database construction method to obtain structured data of the model. Optionally, the data annotation method can be manual annotation and verification, or annotation can be generated based on a large model and combined with manual verification and modification, etc., which is not limited here.

[0266] In some embodiments, a computer device constructs an initial vehicle knowledge base based on structured data from a modeled scene library, structured data from a modeled project library, and structured data from a modeled constraint library. This can be achieved by: associating the structured data from the modeled scene library, structured data from the modeled project library, and structured data from the modeled constraint library based on storage identifiers; associating the original documents, images, or audio / video files corresponding to the structured data from the modeled scene library, structured data from the modeled project library, and structured data from the modeled constraint library based on file access paths; and obtaining the initial vehicle knowledge base based on the associated data.

[0267] The data association can be established between computer devices in the following ways:

[0268] a. Vehicle usage scenario database data storage

[0269] {

[0270] "sceneId": "xxxxxx",

[0271] "projectId": ["xxxxxx","xxxxxx"],

[0272] "constraintId": ["xxxxxx","xxxxxx"],

[0273] "require": "The vehicle supports 220V power supply",

[0274] "requireSource": "Market Feedback", ...

[0276] }

[0277] b. Project database data storage

[0278] {

[0279] "projectId": "xxxxxx",

[0280] "sceneId": ["xxxxxx","xxxxxx"],

[0281] "constraintId": ["xxxxxx","xxxxxx"],

[0282] "projectName": "Project requiring in-vehicle 220V inverter for xx vehicle model",

[0283] "vehicleModel": "xx",

[0284] "priceRange": "450,000-600,000",

[0285] Background: Market feedback indicates that customers have needs for power supply in vehicles and for parking / camping. ...

[0287] }

[0288] c. Vehicle Product Constraint Database Data Storage

[0289] {

[0290] "constraintId": "xxxxxx",

[0291] "category": "law regulation",

[0292] "source": "ISO-26262 Functional Safety Standard",

[0293] "validityPeriod": "March 19, 2025 - ?"

[0294] "attachment": "file"

[0295] }

[0296] Where: sceneId is the scene ID, projectId is the project ID, and constraintId is the constraint ID.

[0297] Through the above storage identifier association, computer devices can find the corresponding data from the other two databases from vehicle usage scenario data or project data.

[0298] In some embodiments, the computer device performs retrieval enhancement generation and parsing processing on the original data associated with the scene library, item library, and constraint library respectively, to obtain the processed data corresponding to the scene library, item library, and constraint library respectively, and uploads the processed data corresponding to the scene library, item library, and constraint library respectively to the initial vehicle knowledge base. The process of obtaining the vehicle knowledge base can be as follows: Figure 5 As shown, Figure 5This is a schematic diagram illustrating a data entry process provided in an embodiment of this application. For example... Figure 5 As shown, for the original data corresponding to each library, the computer device can perform data cleaning and normalization on the original data to obtain cleaned data; perform text segmentation on the cleaned data to obtain segmented data, and vectorize the segmented data to obtain vector data; extract metadata and keywords from the cleaned data to obtain metadata information; store the vector data in the vector library (corresponding to the aforementioned initial vehicle knowledge base), and directly store the metadata in the vector library to obtain the vehicle knowledge base.

[0299] Optionally, the computer equipment performs text segmentation on the cleaned data to obtain segmented data. This can be done using semantic segmentation (such as using sentence boundaries, line breaks, etc.). Optionally, during the text segmentation process, the token length can be less than 2k, and the sliding window overlap can be 30%, thus preserving the complete semantics of a sentence or paragraph.

[0300] Optionally, the computer device can vectorize the segmented data to obtain vector data. This can be achieved by using an open-source text vectorization model (such as the BAAI General Embedding M3 model, BGE-M3) to convert the segmented data into vector data; where the model output dimension can be 1024.

[0301] In some embodiments, after building a vehicle knowledge base, the computer device also builds a knowledge search interface, namely the knowledge search tool mentioned above.

[0302] Please see Figure 6 , Figure 6 This is a schematic diagram illustrating the construction process of a vehicle knowledge base provided in an embodiment of this application. For example... Figure 6 As shown, the computer equipment can perform data cleaning and normalization on survey data, vehicle usage scenario data, decision-making data, and data from other sources to obtain vehicle usage scenario data; it can also perform data cleaning and normalization on project initiation data, process specification data, project outcome data, and project summary and review data to obtain project case data (i.e., the aforementioned vehicle product project data); and it can perform data cleaning and normalization on regulatory standard data, constraint standard data, economic and cultural data, and technical standard data to obtain vehicle product constraint data; based on the vehicle usage scenario data, project case data, and vehicle product constraint data, the data is uploaded to obtain a knowledge base (i.e., the vehicle knowledge base); and then, a knowledge search interface is constructed.

[0303] By adopting this implementation method, the computer device constructs a vehicle knowledge base by combining three databases (scenario database, project database, and constraint database), which can help improve the knowledge completeness of the vehicle knowledge base and thus improve the accuracy of subsequent searches for information related to vehicle product demand analysis data from the vehicle knowledge base.

[0304] In one alternative implementation, Figure 1 In step S102 of the vehicle product analysis and decision-making method shown, the computer device calls a pre-built knowledge search tool to search for first vehicle usage scenario information associated with the vehicle product demand analysis data from a pre-built vehicle knowledge base. This can be achieved by: calling the pre-built knowledge search tool to perform data augmentation processing on the vehicle product demand analysis data to obtain augmented vehicle product demand analysis data; determining a first set of candidate vehicle usage scenario information associated with the augmented vehicle product demand analysis data from the pre-built vehicle knowledge base; converting the augmented vehicle product demand analysis data into vector data; and determining a second set of candidate vehicle usage scenario information from the first set of candidate vehicle usage scenario information with a similarity greater than a preset similarity threshold from the vector data through a multi-path retrieval recall path; wherein different retrieval recall paths correspond to different recall methods; reordering multiple candidate vehicle usage scenario information in the second set of candidate vehicle usage scenario information; and determining the first vehicle usage scenario information associated with the vehicle product demand analysis data based on the reordered multiple candidate vehicle usage scenario information.

[0305] In some embodiments, the computer device performs data augmentation processing on vehicle product demand analysis data to obtain enhanced vehicle product demand analysis data. This may include: performing query correction and completion processing on the vehicle product demand analysis data to obtain processed vehicle product demand analysis data; extracting entities from the processed vehicle product demand analysis data to obtain entities corresponding to the vehicle product demand analysis data; replacing the extracted entities with synonyms based on a thesaurus to obtain replaced entities; converting the vehicle product demand analysis data into vector data to generate semantically similar query variants; and obtaining the enhanced vehicle product demand analysis data based on these semantically similar query variants. In this way, by improving the quality of the input query, the accuracy and relevance of the final generated content can be improved.

[0306] Optionally, the computer equipment performs query, error correction, and completion processing on the vehicle product demand analysis data to obtain the processed vehicle product demand analysis data. This can be achieved by using the edit distance algorithm (Levenshtein distance) to detect spelling errors and correcting them through dictionary matching.

[0307] For example, a computer device can use an N-gram language model to predict the complete query a user might input and perform contextual semantic completion, such as completing "What should I do if my car of model xx breaks down?" into "What should I do if the motor of my car of model xx won't start?".

[0308] Optionally, the computer equipment can extract entities from the processed vehicle product demand analysis data to obtain the entities corresponding to the vehicle product demand analysis data. This can be done by using a Named Entity Recognition (NER) model, such as a Bidirectional Long Short-Term Memory-Conditional Random Field (BiLSTM-CRF), to extract the entities corresponding to the vehicle product demand analysis data from the processed vehicle product demand analysis data.

[0309] For example, suppose the processed vehicle product demand analysis data is "What should I do if the car of model xx breaks down", and the corresponding entity is "car of model xx".

[0310] Optionally, the computer equipment can convert vehicle product demand analysis data into vector data and generate semantically similar query variants. This can be achieved by using the BGE-M3 vectorization model to convert vehicle product demand analysis data into vector data and generate semantically similar query variants.

[0311] The following example illustrates the process of constructing a user request, using the scenario of a user's car failing to start. The user's original question is: "What should I do if my car (model xx) breaks down?" The enhanced text query is: "What should I do if the motor of my car (model xx) won't start?". The computer device can construct the user request as follows:

[0312] {

[0313] "userId": "xxxxxxxxx",

[0314] "carType": "2025 model xx car",

[0315] Query: "What should I do if the motor of my xx car model won't start?"

[0316] }

[0317] The meanings of each field are explained as follows: User ID (userId): User request ID, associated with the user's question and answer context; Car Type (carType): User's car model (optional condition parameter); Query (query): User request enhanced with Query.

[0318] In some embodiments, the computer device determines a first set of candidate vehicle usage scenario information associated with the enhanced vehicle product demand analysis data from a pre-built vehicle knowledge base. This can be achieved by using a scalar retrieval recall method to determine the first set of candidate vehicle usage scenario information associated with the enhanced vehicle product demand analysis data from the vehicle usage scenario data included in the pre-built vehicle knowledge base.

[0319] In some embodiments, the computer device determines a second set of candidate vehicle usage scenario information from the first set of candidate vehicle usage scenario information, whose similarity to the vector data is greater than a preset similarity threshold, through a multi-path retrieval recall path. This can be achieved by using an inverted index recall method and a semantic recall method to determine the second set of candidate vehicle usage scenario information from the first set of candidate vehicle usage scenario information, whose similarity to the vector data is greater than a preset similarity threshold.

[0320] The inverted index recall uses keyword matching (such as Best Matching 25, BM25) to calculate similarity based on the vector data corresponding to the enhanced vehicle product demand analysis data and the term frequency-inverse document frequency (TF-IDF) of each candidate information, and quickly locates documents by building an inverted index.

[0321] Semantic recall employs semantic vector models (such as Bidirectional Encoder Representations from Transformers, BERT) to map the enhanced vehicle product demand analysis data and candidate information into the same vector space, recalling semantically similar documents through cosine similarity. This method can solve the problem of polysemy in long-tail queries. For example, searching for "220V power supply" can recall content related to "power inverter." Data recall can filter out data most similar to the user's question.

[0322] In this process, computer equipment can merge the results of inverted index recall and semantic recall, remove duplicate documents, and obtain a candidate pool of recall results, namely the second candidate scenario information set.

[0323] In some embodiments, the computer device reorders multiple candidate scenario information in the second candidate scenario information set and determines the first vehicle usage scenario information associated with the vehicle product demand analysis data based on the reordered multiple candidate scenario information. This can be achieved by: extracting vector data corresponding to the enhanced vehicle product demand analysis data and feature information of each candidate scenario information in the second candidate scenario information set, the feature information including user behavior feature information and content quality feature information; scoring each candidate scenario information using a Gradient Boosting Decision Tree (GBDT) based on the feature information to obtain a first score corresponding to each candidate scenario information; and, based on the feature information, using a Cross-Encoder reordering algorithm. The Reranker scores each candidate scenario information to obtain a second score for each candidate scenario information. For each candidate scenario information, a comprehensive score is determined based on the first and second scores. The candidate scenario information is then sorted in descending order of comprehensive score to obtain a re-sorted list of candidate scenario information. From the re-sorted list of candidate scenario information, the top N candidate scenario information are selected as the first vehicle usage scenario information associated with the vehicle product demand analysis data.

[0324] Optionally, the computer device determines the comprehensive score corresponding to the candidate scene information based on the first score and the second score corresponding to the candidate scene information. This can be achieved by performing a weighted summation of the first score and the second score corresponding to the candidate scene information to obtain the comprehensive score corresponding to the candidate scene information.

[0325] Optionally, the computer device may also output first vehicle usage scenario information associated with the vehicle product demand analysis data, such as:

[0326] {

[0327] "content": [{

[0328] "require": "The vehicle supports 220V power supply",

[0329] "requireSource": "Market Feedback",

[0330] "user": "Vehicle owner, passenger",

[0331] "scene": ["Car Owner Camping","Work Office - Commute","Work Office - Client Visit"],

[0332] "Industry": "Featured in high-end models, using multiple mature solutions, with costs reduced to below xx yuan."

[0333] "function": "In-vehicle 220V inverter power supply",

[0334] "solve": "Use of high-power electrical appliances in vehicles",

[0335] "case": "rice cooker, hair dryer, laptop",

[0336] "experience": {

[0337] "score": "7.5",

[0338] "frequency": "1-3 times / week",

[0339] "level": "high",

[0340] "cost": "Buyout (less than xx yuan)"

[0341] },

[0342] "attachment": "file"

[0343] "dataSource": "After-sales feedback"

[0344] }]

[0345] }

[0346] In one alternative implementation, Figure 1In step S102 of the vehicle product analysis and decision-making method shown, the computer device calls a pre-built knowledge search tool to search for first vehicle product item information associated with the vehicle product demand analysis data from a pre-built vehicle knowledge base. This can be achieved by: calling the pre-built knowledge search tool to perform data augmentation processing on the vehicle product demand analysis data to obtain augmented vehicle product demand analysis data; determining a first set of candidate vehicle product item information associated with the augmented vehicle product demand analysis data from the pre-built vehicle knowledge base; converting the augmented vehicle product demand analysis data into vector data; and determining a second set of candidate vehicle product item information from the first set of candidate vehicle product item information with a similarity greater than a preset similarity threshold to the vector data through a multi-path retrieval recall path; wherein different retrieval recall paths correspond to different recall methods; reordering multiple candidate vehicle product item information in the second set of candidate vehicle product item information; and determining the first vehicle product item information associated with the vehicle product demand analysis data based on the reordered multiple candidate vehicle product item information.

[0347] In some embodiments, the computer device performs data augmentation processing on the vehicle product demand analysis data to obtain the augmented vehicle product demand analysis data. For details on this, please refer to the description above, which will not be repeated here.

[0348] In some embodiments, the computer device determines a first set of candidate vehicle product item information associated with the enhanced vehicle product demand analysis data from a pre-built vehicle knowledge base. This can be achieved by using a scalar retrieval recall method to determine the first set of candidate vehicle product item information associated with the enhanced vehicle product demand analysis data from the vehicle product item data included in the pre-built vehicle knowledge base.

[0349] In some embodiments, the computer device determines a second set of candidate vehicle product information from the first set of candidate vehicle product information with a similarity greater than a preset similarity threshold to the vector data through a multi-path retrieval recall path. This can be achieved by using an inverted index recall method and a semantic recall method to determine a second set of candidate vehicle product information from the first set of candidate vehicle product information with a similarity greater than a preset similarity threshold to the vector data.

[0350] In some embodiments, the computer device reorders multiple candidate vehicle product item information in the second candidate vehicle product item information set, and determines the first vehicle product item information associated with the vehicle product demand analysis data based on the reordered multiple candidate vehicle product item information. The relevant description can refer to the aforementioned process of the computer device reordering multiple candidate scenario information in the second candidate scenario information set, and determining the first vehicle usage scenario information associated with the vehicle product demand analysis data based on the reordered multiple candidate scenario information, which will not be repeated here.

[0351] In one optional implementation, the computer device invokes a pre-built knowledge search tool to search for first vehicle product constraint information associated with vehicle product demand analysis data from a pre-built vehicle knowledge base. This includes: invoking the pre-built knowledge search tool to perform data augmentation processing on the vehicle product demand analysis data to obtain augmented vehicle product demand analysis data; determining a first set of candidate vehicle product constraint information associated with the augmented vehicle product demand analysis data from the pre-built vehicle knowledge base; converting the augmented vehicle product demand analysis data into vector data; and, through a multi-path retrieval recall path, determining a second set of candidate vehicle product constraint information from the first set of candidate vehicle product constraint information whose similarity to the vector data is greater than a preset similarity threshold; wherein different retrieval recall paths correspond to different recall methods; reordering multiple candidate vehicle product constraint information in the second set of candidate vehicle product constraint information; and, based on the reordered multiple candidate vehicle product constraint information, determining the first vehicle product constraint information associated with the vehicle product demand analysis data.

[0352] In some embodiments, the computer device performs data augmentation processing on the vehicle product demand analysis data to obtain the augmented vehicle product demand analysis data. For details on this, please refer to the description above, which will not be repeated here.

[0353] In some embodiments, the computer device determines a first set of candidate vehicle product constraint information associated with the enhanced vehicle product demand analysis data from a pre-built vehicle knowledge base. This can be achieved by using a scalar retrieval recall method to determine the first set of candidate vehicle product constraint information associated with the enhanced vehicle product demand analysis data from the vehicle product item data included in the pre-built vehicle knowledge base.

[0354] In some embodiments, the computer device determines a second set of candidate vehicle product constraint information from the first set of candidate vehicle product constraint information, whose similarity to the vector data is greater than a preset similarity threshold, through a multi-path retrieval recall path. This can be achieved by using an inverted index recall method and a semantic recall method to determine a second set of candidate vehicle product constraint information from the first set of candidate vehicle product constraint information, whose similarity to the vector data is greater than a preset similarity threshold.

[0355] In some embodiments, the computer device reorders multiple candidate vehicle product constraint information in the second candidate vehicle product constraint information set, and determines the first vehicle product constraint information associated with the vehicle product demand analysis data based on the reordered multiple candidate vehicle product constraint information. The relevant description can refer to the aforementioned process of the computer device reordering multiple candidate scenario information in the second candidate scenario information set, and determining the first vehicle usage scenario information associated with the vehicle product demand analysis data based on the reordered multiple candidate scenario information, which will not be repeated here.

[0356] Please see Figure 7 , Figure 7 This is a schematic diagram illustrating a knowledge search process provided in an embodiment of this application. For example... Figure 7 As shown, (1) the user (i.e. the target object) can input a dialogue request (i.e., vehicle product demand analysis data), and the computer device can obtain the request data (i.e., vehicle product demand analysis data); (2) the request parameters are determined based on the request data; (3) scalar search is used based on the request parameters to search the vehicle knowledge base (i.e., vehicle product demand analysis data). Figure 7 (4) Perform scalar precise query on the Milvus vector library in the database to obtain the initial candidate information set; (5) Perform text data vectorization on the request data to obtain the request vector; (6) Based on the request vector, perform vector similarity search on the initial candidate information set to obtain the first candidate information set; (7) Use the first candidate information set as the retrieval result set; (8) Use the inverted index recall method on the retrieval result set and filter it by keywords to obtain the first recall result set; (9) Use the semantic recall method on the retrieval result set and filter it by semantic similarity to obtain the second recall result set; The first recall result set and the second recall result set are merged to obtain the recall result set; (10) Based on feature engineering, feature information of each information in the request data and the recall result set is extracted; (11) Based on feature information, each information in the recall result set is weighted, fused and reordered to obtain the reordered information; (12) Based on business rules, the reordered information is adjusted to obtain the final reordering result; (13) The TopN result set is selected from the final reordering result, and the selected TopN result set is used as the information that matches the request data searched from the vehicle knowledge base.

[0357] Using this implementation method, the computer device searches for related information associated with vehicle product demand analysis data from the vehicle knowledge base through multi-path recall and reordering, thereby improving the accuracy of search results.

[0358] Please see Figure 8 , Figure 8 This is a flowchart illustrating another vehicle product analysis and decision-making method provided in an embodiment of this application, which can be executed by a computer device. Figure 8 As shown, the vehicle product analysis and decision-making method may include, but is not limited to, the following steps.

[0359] S801. Construct a scenario library to recreate real-world demand scenarios through multi-dimensional vehicle usage scenarios, a project library to align demand analysis with actual goals, and a constraint library to verify the feasibility of demands.

[0360] In an optional implementation, the relevant description of step S801 can be found in the previous description of the computer device constructing a scenario library for reproducing real demand scenarios through multi-dimensional vehicle usage scenarios, a project library for aligning demand analysis and actual goals, and a constraint library for verifying the feasibility of demands, which will not be repeated here.

[0361] S802. Based on the scenario library, project library, and constraint library, construct a vehicle knowledge base.

[0362] In one optional implementation, the computer device constructs a vehicle knowledge base based on a scene library, an item library, and a constraint library in the following manner: Data modeling and structured annotation are performed on the original data associated with each of the scene library, item library, and constraint library to obtain modeled structured data for the scene library, the item library, and the constraint library; an initial vehicle knowledge base is constructed based on these modeled structured data; wherein the modeled structured data for the scene library, the item library, and the constraint library are associated with each other based on storage identifiers; retrieval enhancement and parsing processing is performed on the original data associated with each of the scene library, item library, and constraint library to obtain processed data corresponding to each of the scene library, item library, and constraint library; and the processed data corresponding to each of the scene library, item library, and constraint library is uploaded to the initial vehicle knowledge base to obtain the vehicle knowledge base.

[0363] S803, Build knowledge search tools, network search tools, and constraint and risk assessment tools.

[0364] Among them, the knowledge search tool is used to search from the pre-built vehicle knowledge base for first vehicle usage scenario information, first vehicle product project information, and first vehicle product constraint information associated with the vehicle product demand analysis data; the network search tool is used to search from the entire network for second vehicle usage scenario information, second vehicle product project information, and second vehicle product constraint information associated with the vehicle product demand analysis data; and the constraint and risk assessment tool is used to determine the vehicle product risk assessment result corresponding to the vehicle product demand analysis data based on multiple vehicle product decision factors that match the vehicle product demand analysis data.

[0365] In some embodiments, the function name of the knowledge search tool can be: search_knowledge_base. Parameters may include: query (string) and request_type (optional string), where query represents search keywords, such as "smart home user scenarios," and request_type specifies the search library type, which may include scene (scenario library), project (project library), or constraint (constraint library); if not specified, the entire library is searched by default. Its function is to retrieve matching vehicle usage scenario data (corresponding to the aforementioned first vehicle usage scenario information), historical project cases (corresponding to the aforementioned first vehicle product project information), or constraints (corresponding to the aforementioned first vehicle product constraint information) from the vehicle knowledge base, and return structured information (such as user requirements, functional cases, and regulatory clauses).

[0366] In some embodiments, the function name of the online search tool can be: search_web. Parameters may include: query (string), representing a web-wide search query, such as "latest data privacy constraints". Its function is to supplement real-time external information (such as regulatory standard updates and industry dynamics) to fill the timeliness gaps in the vehicle knowledge base.

[0367] In some embodiments, the function name for the constraint and risk assessment tool may be: assessment_feasibility.

[0368] Parameters (all dictionary type, must include score, weight, and description) may include: national_policy: support for regulatory standards (0-10 points), law_regulation: compliance with constraint standards (0-10 points), economic: economic environment suitability (0-10 points), culture: cultural acceptance (0-10 points), technology_standard: difficulty of technical implementation (0-10 points), cost_benefit: cost-benefit balance (0-10 points). Its function is: based on the dynamic weighting rules of vehicle product decision factors, with all factor weights summed to 1, calculate the feasibility score (0-10 points) and output a risk conclusion (infeasible (4 points and below) / basically feasible (5-8 points) / feasible (9 points and above), where national_policy=0 or law_regulation=0 or culture=0 or technology_standard=0 directly results in 0 points, indicating infeasibility).

[0369] S804, Prompt for building a large model system; The prompt for the large model system includes the functions of the large model and a list of available tools; The list of available tools includes knowledge search tools, network search tools, and constraint and risk assessment tools; The functions include calling the knowledge search tools, calling the network search tools, and calling the constraint and risk assessment tools.

[0370] In some embodiments, the prompts in the large model system may also include role definition, workflow standardization, output requirements, exception handling mechanism, example interaction, etc.

[0371] S805, Process specifications for generating corresponding results from large model construction.

[0372] In some embodiments, the process specifications for generating large model results constructed by computer devices can be found in the description above, and will not be repeated here.

[0373] S806. Based on the prompts and corresponding process specifications for generating large model results, construct a large model for vehicle product analysis and decision-making.

[0374] S807. Obtain the vehicle product demand analysis data input by the target object.

[0375] In an optional implementation, the relevant description of step S807 can be found in the description of step S101 above, and will not be repeated here.

[0376] S808. Based on a pre-built large model for vehicle product analysis and decision-making, a pre-built knowledge search tool is invoked to search the vehicle knowledge base for first vehicle usage scenario information, first vehicle product project information, and first vehicle product constraint information that are related to the vehicle product demand analysis data.

[0377] In one optional implementation, the computer device invokes a pre-built knowledge search tool to search for first vehicle usage scenario information, first vehicle product project information, and first vehicle product constraint information associated with the vehicle product demand analysis data from a pre-built vehicle knowledge base. This can be achieved by: invoking the pre-built knowledge search tool to perform data augmentation processing on the vehicle product demand analysis data, obtaining augmented vehicle product demand analysis data; determining, from the pre-built vehicle knowledge base, a set of first candidate vehicle usage scenario information, a set of first candidate vehicle product project information, and a set of first candidate vehicle product constraint information associated with the augmented vehicle product demand analysis data; converting the augmented vehicle product demand analysis data into vector data; and, through a multi-path retrieval recall path, determining the similarity between the vector data and the first candidate vehicle usage scenario information set, the first candidate vehicle product project information set, and the first candidate vehicle product constraint information set, respectively. The system comprises three sets of information: a second set of candidate vehicle usage scenarios with similarity exceeding a preset threshold, a second set of candidate vehicle product project information, and a second set of candidate vehicle product constraint information. Different retrieval paths correspond to different retrieval methods. The system reorders multiple candidate vehicle usage scenario information items in the second set of information, and based on the reordered multiple candidate vehicle usage scenario information, determines the first vehicle usage scenario information associated with the vehicle product demand analysis data. It also reorders multiple candidate vehicle product project information items in the second set of information, and based on the reordered multiple candidate vehicle product project information, determines the first vehicle product project information associated with the vehicle product demand analysis data. Finally, it reorders multiple candidate vehicle product constraint information items in the second set of information, and based on the reordered multiple candidate vehicle product constraint information, determines the first vehicle product constraint information associated with the vehicle product demand analysis data.

[0378] S809. Call the pre-built network search tool to search the entire network for second vehicle usage scenario information, second vehicle product project information, and second vehicle product constraint information that are related to the vehicle product demand analysis data.

[0379] S810. Extract multiple vehicle product decision factors that match the vehicle product demand analysis data from the first vehicle product constraint information and the second vehicle product constraint information.

[0380] S811. Call the pre-built constraint and risk assessment tool to determine the vehicle product risk assessment result corresponding to the vehicle product demand analysis data based on multiple vehicle product decision factors.

[0381] S812. Based on the first vehicle usage scenario information, the first vehicle product project information, the second vehicle usage scenario information, the second vehicle product project information, and the vehicle product risk assessment results, generate the vehicle product analysis decision results corresponding to the vehicle product demand analysis data.

[0382] In one optional implementation, the computer device generates vehicle product analysis decision results corresponding to vehicle product demand analysis data based on first vehicle usage scenario information, first vehicle product project information, second vehicle usage scenario information, second vehicle product project information, and vehicle product risk assessment results. This can be achieved by generating corresponding process specifications through large model results, and generating vehicle product analysis decision results corresponding to vehicle product demand analysis data based on first vehicle usage scenario information, first vehicle product project information, second vehicle usage scenario information, second vehicle product project information, and vehicle product risk assessment results.

[0383] In this embodiment, on the one hand, since the pre-built vehicle knowledge base includes vehicle usage scenario data for reconstructing real demand scenarios through multi-dimensional vehicle usage scenarios, vehicle product project data for aligning demand analysis with actual goals, and vehicle product constraint data for verifying demand feasibility, using the pre-built vehicle knowledge base to identify and constrain vehicle product demand analysis data can not only provide a basis for user demand analysis but also avoid policy and regulatory risks, thereby improving the accuracy of subsequent vehicle product analysis decision results. On the other hand, by calling pre-built knowledge search tools and pre-built constraint and risk assessment tools for related processing, errors and illusions in the analysis and summary process of the large model used for vehicle product analysis decision-making can be reduced, thereby improving the accuracy of vehicle product analysis decision results.

[0384] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0385] Based on the same inventive concept, this application also provides a vehicle product analysis and decision-making device for implementing the vehicle product analysis and decision-making method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in one or more vehicle product analysis and decision-making device embodiments provided below can be found in the limitations of the vehicle product analysis and decision-making method described above, and will not be repeated here.

[0386] Please see Figure 9 , Figure 9 This is a schematic diagram of the structure of a vehicle product analysis and decision-making device provided in an embodiment of this application. Figure 9 As shown, the vehicle product analysis and decision-making device may include, but is not limited to:

[0387] The acquisition module 901 is used to acquire the vehicle product demand analysis data input by the target object;

[0388] Processing module 902 is used to, based on a pre-built large model for vehicle product analysis and decision-making, invoke a pre-built knowledge search tool to search for first vehicle usage scenario information, first vehicle product project information, and first vehicle product constraint information related to vehicle product demand analysis data from a pre-built vehicle knowledge base. The vehicle knowledge base includes vehicle usage scenario data for reconstructing real demand scenarios through multi-dimensional vehicle usage scenarios, vehicle product project data for aligning demand analysis with actual goals, and vehicle product constraint data for verifying the feasibility of demands. The first vehicle usage scenario information is obtained from the vehicle usage scenario data, the first vehicle product project information is obtained from the vehicle product project data, and the first vehicle product constraint information is obtained from the vehicle product constraint data.

[0389] The determination module 903, based on the first vehicle product constraint information, determines multiple vehicle product decision factors that match the vehicle product demand analysis data, and calls a pre-built constraint and risk assessment tool to determine the vehicle product risk assessment result corresponding to the vehicle product demand analysis data based on the multiple vehicle product decision factors.

[0390] The generation module 904 is used to generate vehicle product analysis and decision results corresponding to the vehicle product demand analysis data based on the first vehicle usage scenario information, the first vehicle product project information, and the vehicle product risk assessment results.

[0391] In some embodiments, the processing module 902 is further configured to invoke a pre-built network search tool to search the entire network for second vehicle usage scenario information, second vehicle product project information, and second vehicle product constraint information associated with the vehicle product demand analysis data; when the determining module 903 determines multiple vehicle product decision factors matching the vehicle product demand analysis data based on the first vehicle product constraint information, it is specifically configured to: extract multiple vehicle product decision factors matching the vehicle product demand analysis data from the first vehicle product constraint information and the second vehicle product constraint information; when the generating module 904 generates vehicle product analysis decision results corresponding to the vehicle product demand analysis data based on the first vehicle usage scenario information, the first vehicle product project information, and the vehicle product risk assessment results, it is specifically configured to: generate vehicle product analysis decision results corresponding to the vehicle product demand analysis data based on the first vehicle usage scenario information, the first vehicle product project information, the second vehicle usage scenario information, the second vehicle product project information, and the vehicle product risk assessment results.

[0392] In some embodiments, the vehicle product risk assessment result includes the target vehicle product risk assessment score corresponding to the vehicle product demand analysis data; when determining the vehicle product risk assessment result corresponding to the vehicle product demand analysis data based on multiple vehicle product decision factors, the determining module 903 is specifically used to: for each vehicle product decision factor among the multiple vehicle product decision factors, obtain the target weight and vehicle product risk assessment score corresponding to the vehicle product decision factor; the target weight is preset or determined based on a predetermined weight allocation strategy; and determine the vehicle product risk assessment result corresponding to the vehicle product demand analysis data based on the target weight and vehicle product risk assessment score corresponding to each vehicle product decision factor.

[0393] In some embodiments, the determining module 903 is further configured to: determine the category to which the vehicle product decision factor belongs and the decision level to which the vehicle product decision factor belongs based on the attributes corresponding to the vehicle product decision factor; determine the initial weight of the vehicle product decision factor based on the decision level; and determine the weight correction coefficient corresponding to the vehicle product decision factor based on the category to which the vehicle product decision factor belongs and the vehicle product risk assessment score corresponding to the vehicle product decision factor; and correct the initial weight based on the weight correction coefficient to obtain the target weight corresponding to the vehicle product decision factor.

[0394] In some embodiments, the plurality of vehicle product decision factors includes at least two of a first vehicle product decision factor, a second vehicle product decision factor, a third vehicle product decision factor, a fourth vehicle product decision factor, a fifth vehicle product decision factor, and a sixth vehicle product decision factor; the first vehicle product decision factor is used to conduct risk assessment on vehicle product demand analysis data from a first dimension, the first dimension representing the regulatory standard support status of the vehicle product corresponding to the vehicle product demand analysis data in the target region; the second vehicle product decision factor is used to conduct risk assessment on vehicle product demand analysis data from a second dimension, the second dimension representing the constraint standard constraint status related to the vehicle product; the third vehicle product decision factor is used to conduct risk assessment on vehicle product demand analysis data from a third dimension, the third dimension representing the economic development status of the region where the vehicle product is used; the fourth vehicle product decision factor is used to conduct risk assessment on vehicle product demand analysis data from a fourth dimension, the fourth dimension representing the cultural situation of the region where the vehicle product is used; the fifth vehicle product decision factor is used to conduct risk assessment on vehicle product demand analysis data from a fifth dimension, the fifth dimension representing the technical standards and parameters required for the realization of the vehicle product; and the sixth vehicle product decision factor is used to conduct risk assessment on vehicle product demand analysis data from a sixth dimension, the sixth dimension representing the cost and benefits required for the realization of the vehicle product.

[0395] In some embodiments, the device may further include a construction module. This construction module is used to construct large model system prompts; the large model system prompts include a list of functions and available tools for the large model; the list of available tools includes pre-built knowledge search tools, network search tools, and constraint and risk assessment tools; the functions include invoking the knowledge search tools, invoking the network search tools, and invoking the constraint and risk assessment tools; wherein, the knowledge search tools are used to search from a pre-built vehicle knowledge base for first vehicle usage scenario information, first vehicle product project information, and first vehicle product constraint information associated with the vehicle product demand analysis data; the network search tools are used to search from the entire network for second vehicle usage scenario information, second vehicle product project information, and second vehicle product constraint information associated with the vehicle product demand analysis data; the constraint and risk assessment tools are used to determine the vehicle product risk assessment result corresponding to the vehicle product demand analysis data based on multiple vehicle product decision factors matching the vehicle product demand analysis data; and based on the large model system prompts, a large model for making vehicle product analysis decisions is constructed.

[0396] In some embodiments, the construction module is further configured to construct a scenario library for recreating real-world demand scenarios through multi-dimensional vehicle usage scenarios, a project library for aligning demand analysis with actual goals, and a constraint library for verifying demand feasibility; perform data modeling and structured annotation on the original data associated with the scenario library, project library, and constraint library respectively, to obtain modeled structured data of the scenario library, modeled structured data of the project library, and modeled structured data of the constraint library; construct an initial vehicle knowledge base based on the modeled structured data of the scenario library, modeled structured data of the project library, and modeled structured data of the constraint library; wherein the modeled structured data of the scenario library, modeled structured data of the project library, and modeled structured data of the constraint library in the vehicle knowledge base are associated based on storage identifiers; perform retrieval enhancement generation and parsing processing on the original data associated with the scenario library, project library, and constraint library respectively, to obtain the processed data corresponding to the scenario library, project library, and constraint library respectively, and upload the processed data corresponding to the scenario library, project library, and constraint library respectively to the initial vehicle knowledge base to obtain the vehicle knowledge base.

[0397] In some embodiments, when the processing module 902 invokes a pre-built knowledge search tool to search for first vehicle usage scenario information associated with vehicle product demand analysis data from a pre-built vehicle knowledge base, it specifically performs the following steps: invokes the pre-built knowledge search tool to perform data augmentation processing on the vehicle product demand analysis data to obtain augmented vehicle product demand analysis data, and determines a first set of candidate vehicle usage scenario information associated with the augmented vehicle product demand analysis data from the pre-built vehicle knowledge base; converts the augmented vehicle product demand analysis data into vector data, and determines a second set of candidate vehicle usage scenario information from the first set of candidate vehicle usage scenario information with a similarity greater than a preset similarity threshold to the vector data through a multi-path retrieval recall path; wherein different retrieval recall paths correspond to different recall methods; reorders multiple candidate vehicle usage scenario information in the second set of candidate vehicle usage scenario information, and determines the first vehicle usage scenario information associated with the vehicle product demand analysis data based on the reordered multiple candidate vehicle usage scenario information.

[0398] In some embodiments, when the processing module 902 invokes a pre-built knowledge search tool to search for first vehicle product item information associated with vehicle product demand analysis data from a pre-built vehicle knowledge base, it specifically performs the following steps: invokes the pre-built knowledge search tool to perform data augmentation processing on the vehicle product demand analysis data to obtain augmented vehicle product demand analysis data, and determines a first set of candidate vehicle product item information associated with the augmented vehicle product demand analysis data from the pre-built vehicle knowledge base; converts the augmented vehicle product demand analysis data into vector data, and determines a second set of candidate vehicle product item information from the first set of candidate vehicle product item information with a similarity greater than a preset similarity threshold to the vector data through a multi-path retrieval recall path; wherein different retrieval recall paths correspond to different recall methods; reorders multiple candidate vehicle product item information in the second set of candidate vehicle product item information, and determines the first vehicle product item information associated with the vehicle product demand analysis data based on the reordered multiple candidate vehicle product item information.

[0399] In some embodiments, when the processing module 902 invokes a pre-built knowledge search tool to search for first vehicle product constraint information associated with vehicle product demand analysis data from a pre-built vehicle knowledge base, it specifically performs the following steps: invokes the pre-built knowledge search tool to perform data augmentation processing on the vehicle product demand analysis data to obtain augmented vehicle product demand analysis data, and determines a first set of candidate vehicle product constraint information associated with the augmented vehicle product demand analysis data from the pre-built vehicle knowledge base; converts the augmented vehicle product demand analysis data into vector data, and determines a second set of candidate vehicle product constraint information from the first set of candidate vehicle product constraint information whose similarity to the vector data is greater than a preset similarity threshold through a multi-path retrieval recall path; wherein different retrieval recall paths correspond to different recall methods; reorders multiple candidate vehicle product constraint information in the second set of candidate vehicle product constraint information, and determines the first vehicle product constraint information associated with the vehicle product demand analysis data based on the reordered multiple candidate vehicle product constraint information.

[0400] It is understood that the specific implementation of each module in the vehicle product analysis and decision-making device provided in this application embodiment and the beneficial effects that can be achieved can be referred to the description of the aforementioned vehicle product analysis and decision-making method embodiment, and will not be repeated here.

[0401] Each module in the aforementioned vehicle product analysis and decision-making device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of the vehicle control device as software, so that the processor can call and execute the corresponding operations of each module.

[0402] In one exemplary embodiment, a computer device is provided, the internal structure of which can be as shown in the figure. Figure 10As shown, the computer device includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The input / output interface is used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements a vehicle product analysis and decision-making method. The display unit is used to form a visually visible image and can be a display screen, projection device, or virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads installed inside the computer device.

[0403] Those skilled in the art will understand that Figure 10 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0404] In one exemplary embodiment, this application provides a computer device including a memory and a processor, wherein the memory stores a computer program; when the processor executes the computer program, it implements the steps in the above-described vehicle product analysis and decision-making methods.

[0405] In one exemplary embodiment, this application provides a computer-readable storage medium having a computer program stored thereon. When executed by a processor, the computer program implements the steps in the above-described vehicle product analysis and decision-making methods.

[0406] In one exemplary embodiment, this application provides a computer program product, including a computer program. When executed by a processor, the computer program implements the steps in the vehicle product analysis and decision-making methods described above.

[0407] Optionally, the computer equipment mentioned in this application can be a terminal or a server. The terminals mentioned herein may include, but are not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices may include smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, projection devices, etc. Portable wearable devices may include smartwatches, smart bracelets, head-mounted devices, etc. Head-mounted devices may include virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, etc. The servers mentioned herein may be independent physical servers, server clusters or distributed systems composed of multiple physical servers, or cloud servers providing cloud computing services, etc., without limitation.

[0408] It should be noted that the data involved in this application (including but not limited to vehicle product demand analysis data, first vehicle usage scenario information, first vehicle product project information, first vehicle product constraint information, vehicle product risk assessment results, etc.) are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0409] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0410] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0411] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A vehicle product analysis and decision-making method, characterized in that, The method includes: Obtain vehicle product demand analysis data input by the target object; Based on a pre-built large model for vehicle product analysis and decision-making, a pre-built knowledge search tool is invoked to search a pre-built vehicle knowledge base for first vehicle usage scenario information, first vehicle product project information, and first vehicle product constraint information associated with the vehicle product demand analysis data. The vehicle knowledge base includes vehicle usage scenario data for reconstructing real demand scenarios through multi-dimensional vehicle usage scenarios, vehicle product project data for aligning demand analysis with actual goals, and vehicle product constraint data for verifying the feasibility of demands. The first vehicle usage scenario information is obtained by searching the vehicle usage scenario data, the first vehicle product project information is obtained by searching the vehicle product project data, and the first vehicle product constraint information is obtained by searching the vehicle product constraint data. Based on the first vehicle product constraint information, multiple vehicle product decision factors that match the vehicle product demand analysis data are determined, and a pre-built constraint and risk assessment tool is invoked to determine the vehicle product risk assessment result corresponding to the vehicle product demand analysis data based on the multiple vehicle product decision factors. Based on the first vehicle usage scenario information, the first vehicle product project information, and the vehicle product risk assessment results, a vehicle product analysis decision result corresponding to the vehicle product demand analysis data is generated.

2. The method according to claim 1, characterized in that, The method further includes: Invoke a pre-built online search tool to search the entire network for second vehicle usage scenario information, second vehicle product project information, and second vehicle product constraint information that are related to the vehicle product demand analysis data; The step of determining multiple vehicle product decision factors that match the vehicle product demand analysis data based on the first vehicle product constraint information includes: From the first vehicle product constraint information and the second vehicle product constraint information, extract multiple vehicle product decision factors that match the vehicle product demand analysis data; The step of generating vehicle product analysis and decision results corresponding to the vehicle product demand analysis data based on the first vehicle usage scenario information, the first vehicle product project information, and the vehicle product risk assessment results includes: Based on the first vehicle usage scenario information, the first vehicle product project information, the second vehicle usage scenario information, the second vehicle product project information, and the vehicle product risk assessment results, a vehicle product analysis decision result corresponding to the vehicle product demand analysis data is generated.

3. The method according to claim 1, characterized in that, The vehicle product risk assessment results include the target vehicle product risk assessment score corresponding to the vehicle product demand analysis data. The process of determining the vehicle product risk assessment result corresponding to the vehicle product demand analysis data based on multiple vehicle product decision factors includes: For each of the multiple vehicle product decision factors, obtain the target weight and risk assessment score corresponding to the vehicle product decision factor; the target weight is preset or determined based on a predetermined weight allocation strategy. Based on the target weights corresponding to each vehicle product decision factor and the vehicle product risk assessment scores, the vehicle product risk assessment results corresponding to the vehicle product demand analysis data are determined.

4. The method according to claim 3, characterized in that, The method further includes: Based on the attributes corresponding to the vehicle product decision factors, determine the category to which the vehicle product decision factors belong, and determine the decision level corresponding to the vehicle product decision factors; Based on the decision-making hierarchy, the initial weights of the vehicle product decision factors are determined, and the weight correction coefficients corresponding to the vehicle product decision factors are determined based on the category to which the vehicle product decision factors belong and the vehicle product risk assessment scores corresponding to the vehicle product decision factors. The initial weights are corrected based on the weight correction coefficients to obtain the target weights corresponding to the vehicle product decision factors.

5. The method according to claim 1, characterized in that, The plurality of vehicle product decision factors include at least two of the following: a first vehicle product decision factor, a second vehicle product decision factor, a third vehicle product decision factor, a fourth vehicle product decision factor, a fifth vehicle product decision factor, and a sixth vehicle product decision factor. The first vehicle product decision factor is used to conduct risk assessment on the vehicle product demand analysis data from a first dimension, whereby the first dimension represents the regulatory standard support status of the vehicle product corresponding to the vehicle product demand analysis data in the target region. The second vehicle product decision factor is used to conduct risk assessment on the vehicle product demand analysis data from a second dimension, whereby the second dimension represents the constraint criteria related to the vehicle product. The third vehicle product decision factor is used to conduct risk assessment on the vehicle product demand analysis data from a third dimension, whereby the third dimension represents the economic development of the region where the vehicle product is used. The fourth vehicle product decision factor is used to conduct risk assessment on the vehicle product demand analysis data from a fourth dimension, and the fourth dimension represents the cultural situation of the region where the vehicle product is used. The fifth vehicle product decision factor is used to conduct risk assessment on the vehicle product demand analysis data from the fifth dimension. The fifth dimension represents the technical standards and parameters required for the realization of the vehicle product. The sixth vehicle product decision factor is used to conduct risk assessment on the vehicle product demand analysis data from a sixth dimension, which represents the cost and benefits required to realize the vehicle product.

6. The method according to any one of claims 1 to 5, characterized in that, The method further includes: The system prompts for building a large model include the functions and a list of available tools for the large model. The list of available tools includes pre-built knowledge search tools, online search tools, and constraint and risk assessment tools. The functions include calling the knowledge search tools, calling the online search tools, and calling the constraint and risk assessment tools. The knowledge search tool is used to search from a pre-built vehicle knowledge base for first vehicle usage scenario information, first vehicle product project information, and first vehicle product constraint information associated with the vehicle product demand analysis data; the network search tool is used to search from the entire network for second vehicle usage scenario information, second vehicle product project information, and second vehicle product constraint information associated with the vehicle product demand analysis data; and the constraint and risk assessment tool is used to determine the vehicle product risk assessment result corresponding to the vehicle product demand analysis data based on multiple vehicle product decision factors that match the vehicle product demand analysis data. Based on the prompts from the large model system, a large model is constructed for vehicle product analysis and decision-making.

7. The method according to claim 6, characterized in that, The method further includes: Build a vehicle usage scenario library to recreate real demand scenarios through multi-dimensional vehicle usage scenarios, a vehicle product project library to align demand analysis with actual goals, and a vehicle product constraint library to verify the feasibility of demands. Data modeling and structured annotation are performed on the original data associated with the vehicle usage scenario library, the vehicle product project library, and the vehicle product constraint library, respectively, to obtain the modeled structured data of the vehicle usage scenario library, the modeled structured data of the vehicle product project library, and the modeled structured data of the vehicle product constraint library. Based on the structured data of the modeled vehicle usage scenario library, the structured data of the modeled vehicle product project library, and the structured data of the modeled vehicle product constraint library, an initial vehicle knowledge base is constructed; wherein, the structured data of the modeled vehicle usage scenario library, the structured data of the modeled vehicle product project library, and the structured data of the modeled vehicle product constraint library in the vehicle knowledge base are associated based on storage identifiers. The original data associated with the vehicle usage scenario library, the vehicle product project library, and the vehicle product constraint library are all subjected to retrieval, enhancement, generation, and parsing processing to obtain the processed data corresponding to the vehicle usage scenario library, the vehicle product project library, and the vehicle product constraint library, respectively. The processed data corresponding to the vehicle usage scenario library, the vehicle product project library, and the vehicle product constraint library are then uploaded to the initial vehicle knowledge base to obtain the vehicle knowledge base.

8. The method according to any one of claims 1 to 5, characterized in that, The step of invoking a pre-built knowledge search tool to search for first vehicle usage scenario information related to the vehicle product demand analysis data from a pre-built vehicle knowledge base includes: The pre-built knowledge search tool is invoked to perform data augmentation processing on the vehicle product demand analysis data to obtain enhanced vehicle product demand analysis data, and a first candidate vehicle usage scenario information set associated with the enhanced vehicle product demand analysis data is determined from the pre-built vehicle knowledge base. The enhanced vehicle product demand analysis data is converted into vector data, and a second set of candidate vehicle usage scenario information with a similarity greater than a preset similarity threshold is determined from the first set of candidate vehicle usage scenario information through a multi-path retrieval recall path; wherein, different retrieval recall paths correspond to different recall methods. The multiple candidate vehicle usage scenario information in the second candidate vehicle usage scenario information set are reordered, and based on the reordered multiple candidate vehicle usage scenario information, the first vehicle usage scenario information associated with the vehicle product demand analysis data is determined.

9. The method according to any one of claims 1 to 5, characterized in that, The step of invoking a pre-built knowledge search tool to search for first vehicle product project information related to the vehicle product demand analysis data from a pre-built vehicle knowledge base includes: The pre-built knowledge search tool is invoked to perform data augmentation processing on the vehicle product demand analysis data to obtain enhanced vehicle product demand analysis data, and a first set of candidate vehicle product project information associated with the enhanced vehicle product demand analysis data is determined from the pre-built vehicle knowledge base. The enhanced vehicle product demand analysis data is converted into vector data, and a second set of candidate vehicle product information with a similarity greater than a preset similarity threshold is determined from the first set of candidate vehicle product information through a multi-path retrieval recall path; wherein, different retrieval recall paths correspond to different recall methods. The multiple candidate vehicle product project information in the second candidate vehicle product project information set are reordered, and based on the reordered multiple candidate vehicle product project information, the first vehicle product project information associated with the vehicle product demand analysis data is determined.

10. The method according to any one of claims 1 to 5, characterized in that, The step of invoking a pre-built knowledge search tool to search for first vehicle product constraint information associated with the vehicle product demand analysis data from a pre-built vehicle knowledge base includes: The pre-built knowledge search tool is invoked to perform data augmentation processing on the vehicle product demand analysis data to obtain enhanced vehicle product demand analysis data, and a first candidate vehicle product constraint information set associated with the enhanced vehicle product demand analysis data is determined from the pre-built vehicle knowledge base. The enhanced vehicle product demand analysis data is converted into vector data, and through a multi-path retrieval recall path, a second set of candidate vehicle product constraint information with a similarity greater than a preset similarity threshold is determined from the first set of candidate vehicle product constraint information; wherein, different retrieval recall paths correspond to different recall methods. The constraint information of multiple candidate vehicle products in the second candidate vehicle product constraint information set is reordered, and based on the reordered constraint information of multiple candidate vehicle products, the first vehicle product constraint information associated with the vehicle product demand analysis data is determined.

11. A vehicle product analysis and decision-making device, characterized in that, The device includes: The acquisition module is used to acquire the vehicle product demand analysis data input by the target object; The processing module is used to, based on a pre-built large model for vehicle product analysis and decision-making, invoke a pre-built knowledge search tool to search a pre-built vehicle knowledge base for first vehicle usage scenario information, first vehicle product project information, and first vehicle product constraint information associated with the vehicle product demand analysis data. The vehicle knowledge base includes vehicle usage scenario data for reconstructing real demand scenarios through multi-dimensional vehicle usage scenarios, vehicle product project data for aligning demand analysis with actual goals, and vehicle product constraint data for verifying demand feasibility. The first vehicle usage scenario information is obtained from the vehicle usage scenario data, the first vehicle product project information is obtained from the vehicle product project data, and the first vehicle product constraint information is obtained from the vehicle product constraint data. The determination module, based on the first vehicle product constraint information, determines multiple vehicle product decision factors that match the vehicle product demand analysis data, and calls a pre-built constraint and risk assessment tool to determine the vehicle product risk assessment result corresponding to the vehicle product demand analysis data based on the multiple vehicle product decision factors. The generation module is used to generate vehicle product analysis and decision results corresponding to the vehicle product demand analysis data based on the first vehicle usage scenario information, the first vehicle product project information, and the vehicle product risk assessment results.

12. A device, characterized in that, It includes a memory and a processor; the memory stores a computer program, characterized in that the processor executes the computer program to implement the steps of the method according to any one of claims 1 to 10.

Images