Vehicle service recommendation method and device, vehicle and storage medium

Abstract

The invention provides a vehicle service recommendation method and device, a vehicle and a storage medium. The method comprises the steps of obtaining user input information and vehicle state information; splicing the user input information and the vehicle state information into initial information; calculating the basic similarity between the initial information and the original fragment information; determining target fragment information from the original fragment information according to the basic similarity and the initial information; according to the target fragment information and the initial information, a service recommendation list is generated, and the service recommendation list comprises at least one candidate service option; and under the condition that a user confirmation operation on the candidate service options is detected, updating the target fragment information by adopting the candidate service options. According to the method, the multiple pieces of target fragment information similar to the initial information are obtained, the service recommendation list is generated, and the service recommendation list comprises at least one candidate service option, so that the model does not need to be continuously trained through historical data.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicles, and in particular, to a vehicle service recommendation method, a vehicle service recommendation device, a vehicle, and a storage medium. Background Art

[0002] Currently, in the technical field of vehicles, vehicle users can interact with the vehicle through voice, text, etc. The vehicle identifies the vehicle user and understands the user's intention during the interaction process, and recommends vehicle control command services according to the vehicle user's intention. However, there are significant individual differences in the usage habits of different vehicle users, and the vehicle control command services recommended by the vehicle cannot meet the needs of different vehicle users.

[0003] When facing this problem, the technical solution adopted by the related art is: constructing a recommendation model, converting the instructions issued by the vehicle user into vectors and then inputting them into the recommendation model to obtain the recommendation result. The recommendation model continuously trains itself through historical data to adapt to the habits of vehicle users. However, this method requires continuous training of the recommendation model according to historical data, and the perception ability of the recommendation model to the changes in the habits of vehicle users is weak, resulting in the vehicle being unable to quickly and accurately identify the intentions of vehicle users. Summary of the Invention

[0004] One of the purposes of the present invention is to provide a vehicle service recommendation method to solve the problem that in the process of service recommendation by the related art, it is necessary to continuously train the recommendation model according to historical data, and the perception ability of the recommendation model to the changes in the habits of vehicle users is weak, resulting in the vehicle being unable to quickly and accurately identify the intentions of vehicle users; the second purpose is to provide a vehicle service recommendation device; the third purpose is to provide a vehicle; the fourth purpose is to provide a computer-readable storage medium.

[0005] In order to achieve the above purposes, the technical solutions adopted by the present invention are as follows:

[0006] In the first aspect, a vehicle service recommendation method is applied to an in-vehicle terminal. The in-vehicle terminal is deployed with a knowledge base, and the knowledge base has original segment information. The method includes:

[0007] Obtain user input information and vehicle status information;

[0008] Concatenate the user input information and the vehicle status information into initial information;

[0009] Calculate the basic similarity between the initial information and the original segment information;

[0010] Determine target segment information from the original segment information according to the basic similarity and the initial information;

[0011] Generate a service recommendation list according to the target segment information and the initial information, where the service recommendation list includes at least one candidate service option;

[0012] In the case of detecting a user confirmation operation for the candidate service option, update the target segment information with the candidate service option.

[0013] Optionally, the determining the target segment information from the original segment information according to the basic similarity and the initial information includes:

[0014] Confirm N candidate segment information from the original segment information according to the basic similarity;

[0015] Convert the initial information into a first feature vector;

[0016] Convert the candidate segment information into a second feature vector;

[0017] Determine the vector similarity between the second feature vector and the first feature vector;

[0018] Sort the candidate segment information according to the vector similarity to determine a segment sequence;

[0019] Determine the first M candidate segment information in the segment sequence as the target segment information, where M < N.

[0020] Optionally, the target segment information corresponds to an original service option, and the generating a service recommendation list according to the target segment information and the initial information includes:

[0021] Fuse the target segment information and the initial information to generate target splicing information;

[0022] Determine the first semantic similarity between the target splicing information and the target segment information;

[0023] Determine the original service option corresponding to the target segment information with the first semantic similarity greater than a preset first similarity threshold as a candidate service option;

[0024] Combine the candidate service options to generate a service recommendation list.

[0025] Optionally, the method further includes:

[0026] Obtain user correction information when generating the service recommendation list;

[0027] Determine the second semantic similarity between the user correction information and the target segment information;

[0028] Confirm the candidate service option corresponding to the target segment information with the second semantic similarity less than or equal to the preset second similarity threshold as the pending service option;

[0029] Determine the original service option corresponding to the target segment information with the second semantic similarity greater than the preset second similarity threshold as the calibration service option;

[0030] Update the pending service option based on the calibration service option.

[0031] Optionally, before calculating the similarity between the initial information and the original segment information, the method further includes:

[0032] Record the operation record information of the initial information generation;

[0033] Store the operation record information.

[0034] Optionally, when detecting a user confirmation operation for the candidate service option, updating the target segment information by using the candidate service option includes:

[0035] When detecting a user confirmation operation for the candidate service option, fuse the candidate service option and the operation record information to generate target record information;

[0036] Update the target segment information by using the target record information.

[0037] Optionally, after updating the target segment information by using the target record information, the method further includes:

[0038] Delete the operation record information.

[0039] In a second aspect, a vehicle service recommendation device is applied to an in-vehicle terminal, and the in-vehicle terminal is deployed with a knowledge base having original segment information. The vehicle service recommendation device includes:

[0040] An acquisition module, configured to acquire user input information and vehicle status information;

[0041] A splicing module, configured to splice the user input information and the vehicle status information into initial information;

[0042] A calculation module, configured to calculate a basic similarity between the initial information and the original segment information;

[0043] A determination module, configured to determine target segment information from the original segment information according to the basic similarity and the initial information;

[0044] A service generation module, configured to generate a service recommendation list according to the target segment information and the initial information, where the service recommendation list includes at least one candidate service option;

[0045] An update module, configured to update the target segment information with the candidate service option when a user confirmation operation for the candidate service option is detected.

[0046] In a third aspect, a vehicle includes a processor, a memory, and a computer program stored on the memory and capable of running on the processor. When the computer program is executed by the processor, the steps of the vehicle service recommendation method described above are implemented.

[0047] In a fourth aspect, a computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the steps of the vehicle service recommendation method described above are implemented.

[0048] In the embodiments of the present invention, the user input information and the vehicle information are spliced into initial information, and then multiple target segment information similar to the initial information is obtained from the knowledge base. According to the target segment information and the initial information, a service recommendation list including at least one candidate service option is generated. Furthermore, the candidate service options are directly output according to the knowledge base, the user input information, and the vehicle status information for the user to select, making the recommendation process simpler and faster. In the embodiments of the present invention, after the recommended service to be executed by the user is confirmed through the user confirmation operation, the confirmed candidate service option is directly used to update the target segment information, thereby updating the knowledge base in a timely manner, enabling the latest preferences of the user to be quickly captured, without separately collecting historical data, reducing problems such as noise, missing values, outliers, and delays in the historical data collection process, thereby strengthening the perception of changes in vehicle user habits during the service recommendation process, and enabling the vehicle to more accurately identify the intentions of vehicle users. Description of the Drawings

[0049] By reading the detailed description of the preferred embodiments below, various other advantages and benefits will become clear to those of ordinary skill in the art. The drawings are only for the purpose of showing the preferred embodiments and are not considered to be a limitation of the present invention. Moreover, throughout the drawings, the same reference numerals are used to represent the same components. In the drawings:

[0050] Figure 1 is a flowchart of the steps of a vehicle service recommendation method provided by an embodiment of the present invention;

[0051] Figure 2 is Figure 1 a flowchart of step 104 in the vehicle service recommendation method provided by an embodiment of the present invention;

[0052] Figure 3 Yes Figure 1 It is the flowchart of step 105 in the vehicle service recommendation method provided by the embodiment of the present invention;

[0053] Figure 4 It is the schematic structural diagram of the vehicle service recommendation device provided by the embodiment of the present invention. Detailed implementation manners

[0054] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the following will elaborate on each implementation manner of the present invention with reference to the accompanying drawings. However, those of ordinary skill in the art can understand that in each implementation manner of the present invention, many technical details are presented for the readers to better understand the present application. However, even without these technical details and various changes and modifications based on the following implementation manners, the technical solutions claimed in the present application can still be implemented. The following division of each embodiment is for convenience of description and should not constitute any limitation to the specific implementation manner of the present invention. Each embodiment can be combined and cross-referenced with each other on the premise of no contradiction.

[0055] Currently, in the field of vehicle technology, to solve the problem that the vehicle control command service recommended by the vehicle cannot meet the needs of different vehicle users, the related technology constructs a recommendation model, converts the commands issued by the vehicle user into vectors and inputs them into the recommendation model to obtain the recommendation result. The recommendation model continuously trains itself through historical data to adapt to the habits of vehicle users. However, this method requires continuous training of the recommendation model according to historical data, and the perception of the recommendation model for changes in the habits of vehicle users is weak, resulting in the vehicle being unable to quickly and accurately identify the intentions of vehicle users. The present invention is proposed to reduce the training cost of the model and enhance the perception of changes in the habits of vehicle users during the service recommendation process.

[0056] Refer to Figure 1 , which shows the step flowchart of an embodiment of the vehicle service recommendation method of the present invention, applied to an in-vehicle terminal. The in-vehicle terminal is deployed with a knowledge base, and the knowledge base has original fragment information. The vehicle service recommendation method may specifically include the following steps:

[0057] Step 101, obtain user input information and vehicle status information;

[0058] In the embodiment of the present invention, the user input information may include one or more of text, voice, pictures or videos, and the vehicle status information includes information such as the current time, location, weather, temperature inside and outside the vehicle, and driving status.

[0059] Step 102, splice the user input information and the vehicle status information into initial information;

[0060] The user input information and vehicle status information are concatenated. Specifically, the vehicle status information is converted into a feature vector, and natural language technology is used to process the user input information to extract text features, and the extracted text features are also converted into feature vectors. The feature vector converted from the text features and the feature vector converted from the vehicle status information are concatenated, so that the generated initial information helps to better understand the owner's intentions and needs, and retrieve the most relevant fragment information from the knowledge base more accurately. For example, if the owner often performs a certain operation at a specific time period and location, the system will use this information as a recommendation basis to provide more personalized service recommendations.

[0061] Step 103, calculate the basic similarity between the initial information and the original fragment information;

[0062] In the embodiment of the present invention, the initial information and the original fragment information can be input into a model based on word vectors, such as Word2Vec or GloVe, etc., or the initial information and the original fragment information can be input into a model based on statistics, such as TF-IDF or Jaccard similarity coefficient, etc. There is no limitation here. After the initial information and the original fragment information are input into the model for calculating similarity, the basic similarity between the initial information and the original fragment information is obtained.

[0063] Step 104, determine the target fragment information from the original fragment information according to the basic similarity and the initial information.

[0064] After obtaining the basic similarity between the initial information and the original fragment information, find the original fragment information with the top basic similarity, and then select the target fragment information from the original fragment information with the top basic similarity according to the initial information again, so as to further improve the similarity between the selected target fragment information and the initial information.

[0065] Step 105, generate a service recommendation list according to the target fragment information and the initial information, and the service recommendation list includes at least one candidate service option.

[0066] In the embodiment of the present invention, after obtaining the target fragment information, the target fragment information and the initial information can be input into a large language model. The large language model can effectively perform service recommendation through steps such as data collection and preprocessing, user intention understanding, service matching and recommendation, generate a service recommendation list, provide more personalized and accurate recommendation results, and improve user satisfaction and experience.

[0067] In other embodiments, the target fragment information and the initial information can also be input into models such as matrix factorization recommendation models or collaborative filtering recommendation models to generate a service recommendation list, which is not limited here.

[0068] Step 106: When a user confirmation operation for a candidate service option is detected, the target segment information is updated using the candidate service option.

[0069] After generating the service recommendation list, obtain the user's selection of candidate service options in the service recommendation list. When it is detected that the user has selected one of the candidate service options, detect the confirmation of the candidate service option. When the user confirmation operation on the candidate service option is detected, the candidate service option is used to update the target fragment information, thereby updating the knowledge base in time so that the user's latest preferences can be captured.

[0070] The embodiment of the present invention concatenates user input information and vehicle information into initial information, then obtains multiple target fragment information similar to the initial information from the knowledge base, and generates a service recommendation list containing at least one candidate service option based on the target fragment information and the initial information, and then directly outputs candidate service options based on the knowledge base, user input information and vehicle status information for user selection, making the recommendation process simpler and faster. After the embodiment of the present invention confirms the recommended service that the user wants to perform later through the user confirmation operation, it directly uses the confirmed candidate service option to update the target fragment information, thereby updating the knowledge base in time, so that the user's latest preferences can be quickly captured without the need to collect historical data separately, reducing the problems of noise, missing values, outliers and delays in the process of collecting historical data, thereby enhancing the perception of changes in vehicle user habits in the service recommendation process, and enabling the vehicle to more accurately identify the vehicle user's intentions.

[0071] In an optional embodiment of the present invention, referring to Figure 2 , determining the target segment information from the original segment information according to the basic similarity and the initial information includes:

[0072] Step 201 : confirm N candidate segment information from the original segment information according to the basic similarity.

[0073] After obtaining the basic similarity between the initial information and the original segment information, find out the top N original segment information with the highest basic similarity, and determine the N original segment information as candidate segment information.

[0074] Step 202, converting the initial information into a first feature vector;

[0075] Step 203: convert the candidate segment information into a second feature vector.

[0076] Step 204: Determine the vector similarity between the second feature vector and the first feature vector.

[0077] Step 205: sort the candidate segment information according to vector similarity to determine a segment sequence.

[0078] After determining the candidate segment information, the initial information is converted into a first feature vector, and the candidate segment information is converted into a second feature vector. By determining the similarity between the feature vectors, the candidate segment information is sorted again according to the similarity, so as to determine the segment sequence.

[0079] In the embodiments of the present application, the similarity between vectors is calculated by a large language model. Specifically, the initial information and the candidate segment information are input into the large language model. The large language model usually uses cosine similarity to calculate the similarity between the embedding vectors of two text information. Since the vector dimension is high and contains rich semantic information, cosine similarity can more accurately reflect the semantic similarity between texts. However, due to its complex model structure and high-dimensional embedding vectors, the large language model has relatively low computational efficiency. And the embodiments of the present application are based on the second similarity calculation of the candidate segment information, and its computational amount is relatively small compared with determining the candidate segment information from the original segment information, which is suitable for the large language model. Therefore, in the process of calculating the similarity for the second time, using the large language model can more accurately sort the candidate segment information.

[0080] Step 206, determine the first M candidate segment information in the segment sequence as the target segment information, where M < N.

[0081] Confirming N candidate segment information from the original segment information according to the basic similarity is not necessarily completely relevant to the target information. By further sorting, the first M candidate segment information in the sorted candidate segment information is used as the target segment information, reducing the irrelevant interference items in the candidate segment information, thus ensuring the similarity between the selected target segment information and the initial information.

[0082] Optionally, in some embodiments, the similarity between vectors can also be calculated by a model based on word vectors and statistics. Specifically, N can be set as an indefinite value, and the original segment information with a basic similarity greater than the preset basic similarity threshold is determined as the candidate segment information by setting the basic similarity threshold. Subsequently, the initial information and the candidate segment information are input into the model based on word vectors and statistics; the model based on word vectors and statistics usually has high computational efficiency, is suitable for processing large-scale data, and is suitable for simple text similarity calculation or scenarios with high requirements for computational efficiency. When the number of candidate segment information confirmed from the original segment information according to the basic similarity threshold is large, the model based on word vectors and statistics can be used to continue the calculation until the number of candidate segment information is less than the preset value, and then the large language model is used to determine the target segment information.

[0083] In an alternative embodiment of the present invention, refer to Figure 3, the target segment information corresponds to an original service option. Generating a service recommendation list based on the target segment information and the initial information includes:

[0084] Step S301, fuse the target segment information and the initial information to generate target splicing information.

[0085] In this embodiment, both the target segment information and the initial information are converted into feature vectors, and the feature vector converted from the target segment information and the feature vector converted from the initial information are fused. Specifically, the feature vector converted from the target segment information and the feature vector converted from the initial information are directly spliced into a high-dimensional vector for fusion; in other embodiments, weights can also be assigned according to the importance of the feature vectors and then weighted fusion can be performed, which is not limited here.

[0086] Step S302, determine the first semantic similarity between the target splicing information and the target segment information.

[0087] After fusing the target segment information and the initial information, determine the similarity between the target splicing information and the target segment information again. Specifically, input the target splicing information and the target segment information into a large language model, and the large language model uses cosine similarity to calculate the similarity between the embedding vectors of the target splicing information and the target segment information, so as to combine the initial information and the target segment information this time, emphasize the role of the user input information and the vehicle status information this time, and generate a service recommendation that better meets the user's needs this time.

[0088] Step S303, determine the original service option corresponding to the target segment information with the first semantic similarity greater than the preset first similarity threshold as the candidate service option.

[0089] Step S304, combine the candidate service options to generate a service recommendation list.

[0090] In this embodiment, the similarity between the target splicing information and the target segment information is calculated, so as to screen the target segment information again while emphasizing the role of the user input information and the vehicle status information this time, improving the perception of changes in vehicle user habits during the service recommendation process, and directly calculating the similarity between the target splicing information and the target segment information, making the calculation amount of the vehicle service recommendation method of this application relatively small, and then generating a service recommendation list more quickly.

[0091] Optionally, in some embodiments, each piece of original segment information corresponds to an original service option. After fusing the target segment information and the initial information to generate the target splicing information, the similarity between the target splicing information and the original segment information can also be determined. The original segment information has a large amount of information content, and the corresponding number of original service options is more. When inputting the target splicing information and the original segment information into the large language model for similarity calculation, more comprehensive candidate service options can be obtained.

[0092] In an alternative embodiment of the present invention, the vehicle service recommendation method further includes:

[0093] Step 107, when generating a service recommendation list, obtain user correction information.

[0094] After generating the service recommendation list, obtain user correction information, so as to determine whether the user has other correction behaviors for the generated service recommendation list.

[0095] Step 108, determine the second semantic similarity between the user correction information and the target segment information.

[0096] After obtaining the user correction information, determine the similarity between the user correction information and the target segment information again, so as to combine the current user correction information, emphasize the role of the current user correction information, and generate a service recommendation that better meets the current needs of the user.

[0097] Step 109, confirm the candidate service options corresponding to the target segment information with the second semantic similarity less than or equal to the preset second similarity threshold as the pending service options.

[0098] Step 1010, determine the original service options corresponding to the target segment information with the second semantic similarity greater than the preset second similarity threshold as the calibration service options.

[0099] Step 1011, update the pending service options based on the calibration service options.

[0100] In this embodiment, after the user receives the service recommendation list, the user judges the candidate service options in the service recommendation list. If the candidate service options in the service recommendation list cannot meet the user's needs, the user continues to input information. The information input by the user again is combined with the target splicing information to generate user correction information, and the second semantic similarity between the user correction information and the target segment information is determined. Thus, the candidate service options corresponding to the target segment information with the second semantic similarity less than or equal to the preset second similarity threshold are replaced to obtain an updated service recommendation list. Subsequently, the user judges the candidate service options in the updated service recommendation list again. If the candidate service options in the service recommendation list still cannot meet the user's needs, the user inputs information according to the needs again, so as to obtain different candidate service options until, after multiple updates, the candidate service options in the service recommendation list meet the user's needs. The multiple corrections and candidate service options during this period are stored in historical_memory. When the vehicle owner confirms the execution, all the candidate service options finally executed after capturing the execution signal are obtained, so as to determine multiple candidate service options required by the user at one time.

[0101] In an alternative embodiment of the present invention, before calculating the similarity between the initial information and the original segment information, the method further includes the following steps:

[0102] Step S401, record the initial information to generate operation record information.

[0103] Step S402, store the operation record information.

[0104] By recording the initial information to generate operation record information and storing the operation record information, the information of each service recommendation is saved.

[0105] In an alternative embodiment of the present invention, when it is detected that there is a user confirmation operation for the candidate service option, updating the target segment information by using the candidate service option includes:

[0106] Step S501, when it is detected that there is a user confirmation operation for the candidate service option, fuse the candidate service option and the operation record information to generate target record information.

[0107] In this embodiment, after detecting a user confirmation operation for the candidate service option, the candidate service option is converted into a feature vector, and the operation record information is also converted into a feature vector. The feature vector converted from the candidate service option and the feature vector converted from the operation record information are fused by splicing. Specifically, the two feature vectors are connected end to end to form a new feature vector, so as to generate target record information.

[0108] Step S502, updating the target segment information by using the target record information.

[0109] In this embodiment, the knowledge base includes a general knowledge base and a preference knowledge base. The general knowledge base is used to store default service options, and the preference knowledge base is used to store user input information and vehicle status information associated with the service options. That is, the original segment information in the knowledge base includes original service options and user input information and vehicle status information corresponding to the original service options. When a user confirmation operation for a candidate service option is detected, the target segment information is updated by using the target record information, that is, the corresponding original segment information is found through the candidate service option, and the user input information and vehicle status information corresponding to the original service option are updated with the operation record information corresponding to the candidate service option.

[0110] When a user confirmation operation for a candidate service option is detected, the target segment information is updated by using the candidate service option, so as to update the knowledge base in a timely manner and capture the user's latest preferences.

[0111] In an optional embodiment of the present invention, after updating the target segment information by using the target record information, the vehicle service recommendation method further includes:

[0112] Step S503, deleting the operation record information.

[0113] If it is detected that the target segment information has been updated by using the target record information, the operation record information is deleted, so as to reduce the influence of the current operation record information on the next service recommendation.

[0114] Refer to Figure 4 , which shows a structural block diagram of an embodiment of a vehicle service recommendation device of the present invention. The vehicle service recommendation device includes:

[0115] An acquisition module 601, configured to acquire user input information and vehicle status information;

[0116] A splicing module 602, configured to splice the user input information and the vehicle status information into initial information;

[0117] A calculation module 603, configured to calculate a basic similarity between the initial information and the original segment information;

[0118] A determination module 604, configured to determine target segment information from the original segment information according to the basic similarity and the initial information;

[0119] A service generation module 605, configured to generate a service recommendation list according to the target segment information and the initial information, where the service recommendation list includes at least one candidate service option;

[0120] An update module 606, configured to update the target segment information with the candidate service option when a user confirmation operation for the candidate service option is detected.

[0121] In an optional embodiment of the present invention, the determination module 604 includes:

[0122] A first determination sub-module, configured to confirm N candidate segment information from the original segment information according to the basic similarity;

[0123] A first conversion module, configured to convert the initial information into a first feature vector;

[0124] A second conversion module, configured to convert the candidate segment information into a second feature vector;

[0125] A second determination sub-module, configured to determine the vector similarity between the second feature vector and the first feature vector;

[0126] A sorting module, configured to sort the candidate segment information according to the vector similarity to determine a segment sequence;

[0127] A third determination sub-module, configured to determine the first M candidate segment information in the segment sequence as the target segment information, where M < N.

[0128] In an optional embodiment of the present invention, the target segment information corresponds to an original service option, and the service generation module 605 further includes:

[0129] A first service generation sub-module, configured to fuse the target segment information and the initial information to generate target splicing information;

[0130] A first similarity confirmation module, configured to determine the first semantic similarity between the target splicing information and the target segment information;

[0131] A first service option determination module, configured to determine the original service option corresponding to the target segment information with the first semantic similarity greater than a preset first similarity threshold as the candidate service option;

[0132] A second service generation sub-module, configured to combine the candidate service options to generate a service recommendation list.

[0133] In an optional embodiment of the present invention, the apparatus further includes:

[0134] A correction module, configured to obtain user correction information when generating the service recommendation list;

[0135] A second similarity confirmation module, configured to determine a second semantic similarity between the user correction information and the target segment information;

[0136] A second service option determination module, configured to confirm the candidate service option corresponding to the target segment information with the second semantic similarity less than or equal to a preset second similarity threshold as a pending service option;

[0137] A third service option determination module, which determines the original service option corresponding to the target segment information with the second semantic similarity greater than the preset second similarity threshold as a calibration service option;

[0138] A calibration module, configured to update the pending service option based on the calibration service option.

[0139] In an alternative embodiment of the present invention, the device further includes:

[0140] A recording module, configured to record the initial information generation operation record information;

[0141] A storage module, configured to store the operation record information.

[0142] In an alternative embodiment of the present invention, the update module 606 further includes:

[0143] A first update sub-module, which, in the case of detecting a user confirmation operation for the candidate service option, fuses the candidate service option and the operation record information to generate target record information;

[0144] A second update sub-module, which updates the target segment information using the target record information.

[0145] In an alternative embodiment of the present invention, the device further includes:

[0146] A deletion module, configured to delete the operation record information after updating the target segment information using the target record information.

[0147] An embodiment of the present invention further provides a vehicle, including:

[0148] A processor, a memory, and a computer program stored on the memory and capable of running on the processor, where when the computer program is executed by the processor, the steps of the vehicle service recommendation method according to any one of the embodiments of the present invention are implemented.

[0149] The vehicle service recommendation method is applied to an in-vehicle terminal, and the in-vehicle terminal is deployed with a knowledge base having original segment information. The method includes:

[0150] Obtain user input information and vehicle status information;

[0151] Concatenate the user input information and the vehicle status information into initial information;

[0152] Calculate the basic similarity between the initial information and the original fragment information;

[0153] Determine target fragment information from the original fragment information according to the basic similarity and the initial information;

[0154] Generate a service recommendation list according to the target fragment information and the initial information, where the service recommendation list includes at least one candidate service option;

[0155] When a user confirmation operation for the candidate service option is detected, update the target fragment information with the candidate service option.

[0156] Optionally, the determining target fragment information from the original fragment information according to the basic similarity and the initial information includes:

[0157] Confirm N candidate fragment information from the original fragment information according to the basic similarity;

[0158] Convert the initial information into a first feature vector;

[0159] Convert the candidate fragment information into a second feature vector;

[0160] Determine the vector similarity between the second feature vector and the first feature vector;

[0161] Sort the candidate fragment information according to the vector similarity to determine a fragment sequence;

[0162] Determine the first M candidate fragment information in the fragment sequence as the target fragment information, where M < N.

[0163] Optionally, the target fragment information corresponds to an original service option, and the generating a service recommendation list according to the target fragment information and the initial information includes:

[0164] Fuse the target fragment information and the initial information to generate target concatenated information;

[0165] Determine the first semantic similarity between the target concatenated information and the target fragment information;

[0166] Determine the original service option corresponding to the target fragment information with the first semantic similarity greater than a preset first similarity threshold as the candidate service option;

[0167] Combine the candidate service options to generate a service recommendation list.

[0168] Optionally, the method further includes:

[0169] When generating the service recommendation list, obtaining user correction information;

[0170] Determining a second semantic similarity between the user correction information and the target segment information;

[0171] Identifying the candidate service options corresponding to the target segment information with the second semantic similarity less than or equal to a preset second similarity threshold as pending service options;

[0172] Determining the original service options corresponding to the target segment information with the second semantic similarity greater than the preset second similarity threshold as calibrated service options;

[0173] Updating the pending service options based on the calibrated service options.

[0174] Optionally, before calculating the similarity between the initial information and the original segment information, the method further includes:

[0175] Recording operation record information of the initial information generation;

[0176] Storing the operation record information.

[0177] Optionally, when detecting a user confirmation operation for the candidate service option, the step of updating the target segment information with the candidate service option includes:

[0178] When detecting a user confirmation operation for the candidate service option, fusing the candidate service option and the operation record information to generate target record information;

[0179] Updating the target segment information with the target record information.

[0180] Optionally, after updating the target segment information with the target record information, the method further includes:

[0181] Deleting the operation record information.

[0182] The above-mentioned memory may include a random access memory (Random Access Memory, RAM for short), or may also include non-volatile memory, such as at least one disk memory. Optionally, the memory may also be at least one storage device located far from the aforementioned processor.

[0183] The above-mentioned processor may be a general-purpose processor, including a Central Processing Unit (CPU for short), a Network Processor (NP for short), etc.; it may also be a Digital Signal Processor (DSP for short), an Application Specific Integrated Circuit (ASIC for short), a Field-Programmable Gate Array (FPGA for short), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0184] An embodiment of the present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the vehicle service recommendation method according to any one of the embodiments of the present invention.

[0185] A vehicle service recommendation method is applied to an in-vehicle terminal. The in-vehicle terminal is deployed with a knowledge base, and the knowledge base has original segment information. The method includes:

[0186] Obtain user input information and vehicle status information;

[0187] Concatenate the user input information and the vehicle status information into initial information;

[0188] Calculate the basic similarity between the initial information and the original segment information;

[0189] Determine target segment information from the original segment information according to the basic similarity and the initial information;

[0190] Generate a service recommendation list according to the target segment information and the initial information. The service recommendation list includes at least one candidate service option;

[0191] When a user confirmation operation for the candidate service option is detected, update the target segment information with the candidate service option.

[0192] Optionally, the determining target segment information from the original segment information according to the basic similarity and the initial information includes:

[0193] Confirm N candidate segment information from the original segment information according to the basic similarity;

[0194] Convert the initial information into a first feature vector;

[0195] Convert the candidate segment information into a second feature vector;

[0196] Determine the vector similarity between the second feature vector and the first feature vector;

[0197] Sort the candidate segment information according to the vector similarity to determine a segment sequence;

[0198] Determine the first M candidate segment information in the segment sequence as the target segment information, where M < N.

[0199] Optionally, the target segment information corresponds to an original service option. Generating a service recommendation list according to the target segment information and the initial information includes:

[0200] Fuse the target segment information and the initial information to generate target splicing information;

[0201] Determine the first semantic similarity between the target splicing information and the target segment information;

[0202] Determine the original service option corresponding to the target segment information with the first semantic similarity greater than a preset first similarity threshold as a candidate service option;

[0203] Combine the candidate service options to generate a service recommendation list.

[0204] Optionally, the method further includes:

[0205] When generating the service recommendation list, obtain user correction information;

[0206] Determine the second semantic similarity between the user correction information and the target segment information;

[0207] Determine the candidate service option corresponding to the target segment information with the second semantic similarity less than or equal to a preset second similarity threshold as a pending service option;

[0208] Determine the original service option corresponding to the target segment information with the second semantic similarity greater than a preset second similarity threshold as a calibrated service option;

[0209] Update the pending service option based on the calibrated service option.

[0210] Optionally, before calculating the similarity between the initial information and the original segment information, the method further includes:

[0211] Record the initial information to generate operation record information;

[0212] Store the operation record information.

[0213] Optionally, when a user confirmation operation for the candidate service option is detected, updating the target segment information by using the candidate service option includes:

[0214] When a user confirmation operation for the candidate service option is detected, fusing the candidate service option and the operation record information to generate target record information;

[0215] Updating the target segment information by using the target record information.

[0216] Optionally, after updating the target segment information by using the target record information, the method further includes:

[0217] Deleting the operation record information.

[0218] Each embodiment in the specification is described in a progressive manner. Each embodiment focuses on the differences from other embodiments. For the same or similar parts among the embodiments, reference may be made to each other.

[0219] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a device, or a computer program product. Therefore, the embodiments of the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the embodiments of the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0220] The embodiments of the present invention are described with reference to the flowcharts and / or block diagrams of the method, terminal device (system), and computer program product according to the embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, and the combination of the processes and / or blocks in the flowchart and / or block diagram can also be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing terminal devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal devices generate a device for implementing the specified functions in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0221] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing terminal device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device, and the instruction device implements the functions specified in the process Figure 1One or more processes and / or blocks Figure 1 The functions specified in one or more blocks.

[0222] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device, so that a series of operation steps are executed on the computer or other programmable terminal device to generate a computer-implemented process. Thus, the instructions executed on the computer or other programmable terminal device provide steps for implementing the functions specified in one Figure 1 One or more processes and / or blocks Figure 1 The steps of the functions specified in one or more blocks.

[0223] The above embodiments are only preferred embodiments given to fully illustrate the present invention, and the protection scope of the present invention is not limited thereto. Equivalent substitutions or transformations made by those skilled in the art on the basis of the present invention are within the protection scope of the present invention.

Claims

1. A vehicle service recommendation method, characterized in that: Applied to a vehicle-mounted terminal, the vehicle-mounted terminal is deployed with a knowledge base, the knowledge base has original fragment information, and the method includes: Obtain user input information and vehicle status information; splicing the user input information and the vehicle status information into initial information; Calculating a basic similarity between the initial information and the original segment information; determining target segment information from the original segment information according to the basic similarity and the initial information; generating a service recommendation list according to the target segment information and the initial information, wherein the service recommendation list includes at least one candidate service option; In a case where a user confirmation operation on the candidate service option is detected, the target segment information is updated using the candidate service option.

2. The method according to claim 1, characterized in that The determining the target segment information from the original segment information according to the basic similarity and the initial information comprises: Confirming N candidate segment information from the original segment information according to the basic similarity; Converting the initial information into a first eigenvector; Converting the candidate segment information into a second feature vector; Determining a vector similarity between the second feature vector and the first feature vector; Sorting the candidate fragment information according to the vector similarity to determine a fragment sequence; The first M candidate fragment information in the fragment sequence is determined as the target fragment information, where M <N。 3. The method according to claim 1, characterized in that The target segment information corresponds to an original service option, and generating a service recommendation list according to the target segment information and the initial information includes: Fusion of the target fragment information and the initial information to generate target splicing information; Determining a first semantic similarity between the target splicing information and the target segment information; Determine the original service option corresponding to the target segment information whose first semantic similarity is greater than a preset first similarity threshold as a candidate service option; The candidate service options are combined to generate a service recommendation list.

4. The method according to claim 3, characterized in that The method further comprises: In the case of generating the service recommendation list, obtaining user correction information; Determining a second semantic similarity between the user correction information and the target segment information; Confirming the candidate service option corresponding to the target segment information whose second semantic similarity is less than or equal to a preset second similarity threshold as a pending service option; Determine the original service option corresponding to the target segment information whose second semantic similarity is greater than a preset second similarity threshold as the calibration service option; The pending service option is updated based on the calibration service option.

5. The method according to claim 1, characterized in that Before calculating the similarity between the initial information and the original segment information, the method further includes: Recording the initial information to generate operation record information; The operation record information is stored.

6. The method according to claim 5, characterized in that When a user confirmation operation on the candidate service option is detected, updating the target segment information using the candidate service option includes: When a user confirmation operation on the candidate service option is detected, fusing the candidate service option and the operation record information to generate target record information; The target segment information is updated using the target record information.

7. The method according to claim 6, characterized in that After the target record information is used to update the target segment information, the method further includes: Delete the operation record information.

8. A vehicle service recommendation device, characterized in that: Applied to a vehicle-mounted terminal, the vehicle-mounted terminal is deployed with a knowledge base, the knowledge base has original fragment information, and the vehicle service recommendation device includes: An acquisition module, used to acquire user input information and vehicle status information; A splicing module, used for splicing the user input information and the vehicle status information into initial information; A calculation module, used for calculating the basic similarity between the initial information and the original segment information; a determination module, configured to determine target segment information from the original segment information according to the basic similarity and the initial information; A service generation module, configured to generate a service recommendation list according to the target segment information and the initial information, wherein the service recommendation list includes at least one candidate service option; The updating module is configured to update the target segment information by using the candidate service option when a user confirmation operation on the candidate service option is detected.

9. A vehicle, characterized in that: The invention comprises a processor, a memory and a computer program stored in the memory and capable of running on the processor, wherein when the computer program is executed by the processor, the steps of the vehicle service recommendation method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the vehicle service recommendation method according to any one of claims 1 to 7 are implemented.

Images