Business suggestion generation method and system based on large model

Abstract

The invention provides a business suggestion generation method and system based on a large model, and the method comprises the steps: S1, obtaining user comment data, and then carrying out the preprocessing of the user comment data, and obtaining training data; s2, a large model is built, then the training data are used for training the large model, a user pain point analysis model is obtained, and the large model comprises the user pain point analysis model and a service suggestion generation model; s3, inputting to-be-analyzed user comment data into the user pain point analysis model to obtain user pain point information; and S4, the business suggestion generation model obtains the market trend information in real time, and then integrates the market trend information with the user pain point information to generate a business suggestion. Due to the fact that business decision makers may have cognitive deviation due to incomplete information acquisition, the network information acquired by the large model in real time can supplement part of missing information of the business decision makers, so that the business decision makers can more comprehensively understand market trends, industry standards, competitor conditions and the like.

Description

Technical Field

[0001] The present application relates to the technical field of natural language processing, and in particular to a method and system for generating business suggestions based on a large model. Background Art

[0002] In today's digital age, e-commerce platforms and various online channels have accumulated a large amount of user review data. These reviews contain rich information such as user experience, opinions and expectations of the product. However, companies and developers currently face many difficulties in obtaining and utilizing this information, and lack effective methods to systematically analyze massive user reviews. Traditional manual analysis methods are inefficient and difficult to fully and accurately grasp user feedback, while some existing automated analysis tools are insufficient in terms of the depth and accuracy of information extraction, the accuracy of sentiment analysis, and the provision of valuable improvement suggestions, and are unable to access large models or generate targeted reports. Summary of the invention

[0003] This application provides a method and system for generating business suggestions based on a large model. The specific technical solution is as follows: A method for generating business suggestions based on a big model, the method comprising the following steps: step S1, obtaining user comment data and business profile data, and then preprocessing the user comment data to obtain training data; step S2, building a big model, and then training the big model with the training data and business profile data to obtain a user pain point analysis model and a business suggestion generation model, wherein the big model comprises the user pain point analysis model and the business suggestion generation model; step S3, inputting the user comment data to be analyzed into the user pain point analysis model to obtain user pain point information; step S4, the business suggestion generation model obtains market trend information in real time, and then integrates the information with the user pain point information to generate business suggestions.

[0004] Furthermore, in the step S1, the method for preprocessing the user comment data to obtain training data includes: step S11, using OCR text recognition technology combined with API interface technology to obtain user comment data from e-commerce platforms, social media and / or product forums; step S12, using regular expressions and string processing functions to clean the user comment data, and then using the jieba Chinese word segmentation library and / or the nltk English word segmentation library to perform word segmentation processing; step S13, annotating the data after word segmentation processing based on a pre-established annotation guide, and then dividing the annotated data into a training set, a validation set and a test set according to a preset ratio, thereby obtaining training data.

[0005] Furthermore, in the step S2, a large model is built, and then the large model is trained using training data and business profile data to obtain a method for obtaining a user pain point analysis model and a business suggestion generation model, including: step S21, selecting a Transformer architecture, and building a large model based on a PyTorch framework; step S22, defining an input layer, and using a pre-trained word vector model to convert user comment data into a word vector of a fixed dimension; step S23, constructing a multi-layer Transformer encoder layer, setting the number of heads of a multi-head attention mechanism and the parameters of a feedforward neural network; step S24, designing an output layer, for the user pain point analysis task, the output layer is designed as a multi-classifier for outputting different types of user pain point information, and for the business suggestion generation task, the output layer is designed as a text generation structure for generating relevant business suggestion texts; step S25, initializing hyperparameters, setting a learning rate, batch size, number of training rounds, and hidden layer dimension; step S26, inputting the training set and business profile data into the large model, and training using an Adam optimizer to obtain a user pain point analysis model and a business suggestion generation model.

[0006] Furthermore, in the step S3, the user comment data to be analyzed is input into the user pain point analysis model, and the method for obtaining user pain point information includes: step S31, cleaning and segmenting the user comment data to be analyzed and converting them into word vectors of fixed dimension, and then inputting them into the user pain point analysis model; step S32, the user pain point analysis model performs semantic association analysis on the words and phrases of the user comment data to be analyzed through a multi-head attention mechanism to understand the user's intention, judges the emotional tendency of the user comment data to be analyzed through sentiment analysis, and discovers common patterns and themes in the user comment data to be analyzed through pattern recognition to obtain user pain point information.

[0007] Furthermore, the step S32 also includes: by calculating the cosine similarity between different user pain point information, the user pain point information with similarity exceeding a threshold is classified into one category, so as to achieve classification and induction of the user pain point information.

[0008] Furthermore, in the step S4, the method in which the business suggestion generation model obtains market trend information in real time and then integrates it with user pain point information to generate business suggestions includes: step S41, the business suggestion generation model uses a time series analysis method and a topic model analysis method to obtain and determine the main themes and hot topics in the market trend information; step S42, counting the frequency of occurrence of different types of user pain point information to determine the priority of the user pain point information; step S43, comparing the user pain point information with the market trend information one by one according to the priority of the user pain point information and outputting the comparison results, and then integrating the comparison results to generate business suggestions.

[0009] Further, in step S43, the method of comparing the user pain point information with the market trend information one by one according to the priority and outputting the comparison results includes: step S431, comparing the potential connection between the main themes and hot topics in the market trend information and the user pain point information, analyzing whether there is an opportunity to use the market trend information to solve the user pain points and outputting relevant results; step S432, analyzing whether there is a conflict between the market trend information and the user pain point information and outputting relevant results.

[0010] Furthermore, in step S43, the method of integrating the comparison results to generate business suggestions includes: analyzing the correlation between market trend information and user pain point information, and then integrating the comparison results between the mutually related market trend information and user pain point information and comprehensively analyzing and elaborating the business suggestions.

[0011] Furthermore, the method for integrating the comparison results between the interrelated market trend information and the user pain point information includes: for text information, using a text fusion algorithm to merge and describe the interrelated market trend information and the user pain point information to generate a coherent text paragraph; for data information, using a data fusion algorithm to fuse the interrelated market trend information and the user pain point information to generate a new data set.

[0012] A big model-based business suggestion generation system is provided, wherein the big model-based business suggestion generation system is used for the big model-based business suggestion generation method.

[0013] The method for generating business recommendations based on a large model described in this application collects user review data to gain an in-depth understanding of the product's performance in the market, user needs and pain points, and the overall state of the industry, and accesses a large language model to summarize and provide business recommendations for massive amounts of user review information. Since business decision makers may have cognitive biases due to incomplete information acquisition, the network information acquired in real time by the large model described in this application can supplement some of the information missing from business decision makers, allowing business decision makers to have a more comprehensive understanding of market trends, industry standards, and competitor situations. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 A flowchart of a method for generating business recommendations based on a large model according to an embodiment of the present application. DETAILED DESCRIPTION

[0015] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.

[0016] It should be understood that when used in this application, the term "comprising" indicates the presence of the described features, integral bodies, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integral bodies, steps, operations, elements, components and / or their collections. It should also be understood that the term "and / or" used in this application refers to any combination of one or more of the associated listed items and all possible combinations, and includes these combinations.

[0017] As used in this application, the term "if" can be interpreted as "when" or "uponce" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" can be interpreted as meaning "uponce it is determined" or "in response to determining" or "uponce [described condition or event] is detected" or "in response to detecting [described condition or event]", depending on the context.

[0018] In addition, in the description of the present application, the terms "first", "second", "third", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance. The reference to "one embodiment" or "some embodiments" described in the specification of the present application means that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment. Therefore, the statements "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in the differences in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0019] In today's digital age, e-commerce platforms and various online channels have accumulated a large amount of user review data. These reviews contain rich information such as user experience, opinions and expectations of the product. However, companies and developers currently face many difficulties in obtaining and utilizing this information, and lack effective methods to systematically analyze massive user reviews. Traditional manual analysis methods are inefficient and difficult to fully and accurately grasp user feedback, while some existing automated analysis tools are insufficient in terms of the depth and accuracy of information extraction, the accuracy of sentiment analysis, and the provision of valuable improvement suggestions, and are unable to access large models or generate targeted reports.

[0020] In order to solve the above technical problems, the embodiment of the present application provides a business suggestion generation method based on a big model. By collecting user review data, it is possible to gain an in-depth understanding of the product's performance in the market, user needs and pain points, and the overall status of the industry, and access a big language model to summarize and make business suggestions for massive user review information.

[0021] like Figure 1 As shown, the business suggestion generation method based on the big model includes the following steps: step S1, obtaining user comment data and business profile data, and then preprocessing the user comment data to obtain training data; step S2, building a big model, and then using the training data and business profile data to train the big model to obtain a user pain point analysis model and a business suggestion generation model, wherein the big model includes a user pain point analysis model and a business suggestion generation model; step S3, inputting the user comment data to be analyzed into the user pain point analysis model to obtain user pain point information; step S4, the business suggestion generation model obtains market trend information in real time, and then integrates it with the user pain point information to generate business suggestions.

[0022] It should be noted that the big model refers to an artificial intelligence big model, which is a machine learning model with ultra-large-scale parameters (usually more than one billion) and complex computing structure. It can process massive data and complete various complex tasks, such as natural language processing, image recognition, etc. The user pain point analysis model and the business suggestion generation model described in this application are divided by function, and are not two independent models. The big model described in this application has both user pain point analysis and business suggestion generation capabilities. It is an organic whole that realizes these two functions. The user pain point analysis model and the business suggestion generation model share knowledge and parameters, and cooperate with each other to generate business suggestions coherently and effectively.

[0023] As one of the implementation modes, in step S1, the method for preprocessing the user comment data to obtain training data includes: step S11, using OCR text recognition technology combined with API interface technology to obtain user comment data from e-commerce platforms, social media and / or product forums; step S12, using regular expressions and string processing functions to clean the user comment data, remove HTML tags and special characters, etc., and then use jieba Chinese word segmentation library and / or nltk English word segmentation library for word segmentation processing, and convert all English texts to lowercase and remove stop words; step S13, based on the pre-established annotation guide, the data after word segmentation processing is annotated, and the definition and distinction criteria of the user pain point information type and the annotation method of the business suggestion related information are clarified. For example, for product quality pain points, it is necessary to clarify whether it is a problem of appearance quality, material quality or other aspects. Then the annotated data is divided into a training set, a validation set and a test set according to a preset ratio of 8:1:1, so as to obtain training data. Among them, data annotation adopts a manual annotation method.

[0024] As one of the implementation modes, in step S2, a large model is built, and then the large model is trained using training data and business data to obtain a user pain point analysis model and a business suggestion generation model. The method includes: step S21, selecting a Transformer architecture, and building a large model based on a PyTorch framework; step S22, defining an input layer, and using a pre-trained word vector model (such as Word2Vec or GloVe) to convert user comment data into a fixed-dimensional word vector; step S23, constructing a multi-layer Transformer encoder layer, and setting the heads of a multi-head attention mechanism. The number and parameters of the feedforward neural network; Step S24, design the output layer. For the user pain point analysis task, the output layer is designed as a multi-classifier to output different types of user pain point information. For the business suggestion generation task, the output layer is designed as a text generation structure to generate relevant business suggestion texts; Step S25, initialize the hyperparameters, set the learning rate to 0.001, the batch size to 32, the number of training rounds to 100, and the hidden layer dimension to 512; Step S26, input the training set and business data into the large model, use the Adam optimizer for training, and obtain the user pain point analysis model and the business suggestion generation model. In the process of training the user pain point analysis model, after a certain number of training rounds (10 rounds are initially planned), calculate the cross entropy loss function (cross entropy loss function is used for classification tasks, and mean square error is used for text generation tasks) and accuracy and other indicators, and adjust the learning rate according to the changes in indicators (using the learning rate decay strategy). After the training is completed, the model performance is evaluated on the validation set. When the accuracy on the validation set no longer increases or begins to decrease, stop the training, use the early stopping method to prevent overfitting, and finally use the test set to verify the performance of the model. In the process of training the business suggestion generation model, in order to ensure that the output expression of the business suggestion generation model is accurate and professional, a large number of network reports, industry white papers and financial reports of listed companies are selected as important training data and input into the big model. These data sources are extensive and authoritative, covering a wealth of industry knowledge and professional expressions. By allowing the big model to deeply learn the content, language style and logical structure, it continuously optimizes its own expression ability, so that it can present business suggestions in an accurate and professional manner in the subsequent output, in line with industry norms and standards. At the same time, a large number of existing business reports of the company can be provided to the big model as key inputs. The company's business report records in detail the various situations of its own business development and has a specific output format and expression habits. By learning these business reports, the big model can be familiar with and master the output format settings suitable for the company's business scenarios, and at the same time make its expression more clear and structured. This helps to output business suggestions in a structured way that is easy to understand and accept within the company when generating business suggestions. In addition, past historical data, that is, previous business decision documents and decision-making processes, can also be input into the big model for training.These historical data carry the company's past practical experience and actual conditions in business decision-making. With the help of learning from these materials, the big model can gain an in-depth understanding of key information such as the company's actual needs, decision-making considerations, etc. in different business scenarios. In this way, when generating business suggestions, it can better combine the company's actual business situation to make the output results more in line with the company's real needs. After sufficient training with the above three different types of business data, the business suggestion generation model described in this application is obtained. When the business suggestion generation model receives the user pain point analysis results output by the user pain point analysis model, it can combine the accurate and professional expression ability, structured output method and in-depth understanding of the company's actual business learned before, so as to accurately output targeted business suggestions that fit the company's actual business scenarios.

[0025] As one of the implementation modes, in the step S3, the user comment data to be analyzed is input into the user pain point analysis model, and the method for obtaining user pain point information includes: step S31, cleaning and segmenting the user comment data to be analyzed and converting them into word vectors of fixed dimension, and then inputting them into the user pain point analysis model; step S32, the user pain point analysis model performs semantic association analysis on the words and phrases of the user comment data to be analyzed through a multi-head attention mechanism to understand the user's intention, judges the emotional tendency of the user comment data to be analyzed through sentiment analysis, and discovers common patterns and themes in the user comment data to be analyzed through pattern recognition to obtain user pain point information.

[0026] As one implementation method, step S32 further includes: calculating the cosine similarity between different user pain point information, classifying the user pain point information with similarity exceeding a threshold into one category, and realizing classification and induction of the user pain point information.

[0027] As one of the implementation modes, in step S4, the business suggestion generation model acquires market trend information in real time, and then integrates it with user pain point information to generate business suggestions, including: step S41, the business suggestion generation model uses time series analysis method and topic model analysis method to observe the change of market trend information over time, find out emerging market trends and development directions, and acquire and determine the main themes and hot topics in the market trend information. Among them, the business suggestion generation model has a networking function, can acquire network information in real time, and connect to relevant information sources on the Internet, such as industry report websites, market research institution databases, news media platforms, social media platforms, etc. by setting a network information acquisition interface. Using application programming interface (API) calling technology, the latest market trend information is acquired in real time, and the acquired information is screened and sorted, irrelevant information and noise data are removed, and key information related to the business is retained. Step S42, the frequency of occurrence of different types of user pain point information is counted to determine the priority of user pain point information; step S43, according to the priority of user pain point information, the information is compared with the market trend information one by one and the comparison results are output, and then the comparison results are integrated to generate business suggestions.

[0028] As one of the implementation modes, in step S43, the method of comparing the user pain point information with the market trend information one by one according to the priority and outputting the comparison results includes: step S431, comparing the potential connection between the main themes and hot topics in the market trend information and the user pain point information, analyzing whether there is an opportunity to use the market trend information to solve the user pain points and outputting relevant results; step S432, analyzing whether there is a conflict between the market trend information and the user pain point information and outputting relevant results.

[0029] As one of the implementation modes, in step S43, the method of integrating the comparison results to generate business suggestions includes: analyzing the correlation between market trend information and user pain point information, and then integrating the comparison results between the mutually related market trend information and user pain point information and comprehensively analyzing and elaborating the business suggestions.

[0030] As one of the implementation modes, the method for integrating the comparison results between the interrelated market trend information and the user pain point information includes: for text information, using a text fusion algorithm to merge and describe the interrelated market trend information and the user pain point information to generate a coherent text paragraph; for data information, using a data fusion algorithm to fuse the interrelated market trend information and the user pain point information to generate a new data set.

[0031] Based on the above implementation method, after outputting the business suggestions, use the front-end development framework (such as Vue.js or React.js) to build a user interface, and present the user pain point analysis results and business suggestions to the user in the form of charts and text reports. Use bar charts and pie charts to show the proportion of different types of user pain points, and use text reports to describe in detail the specific content of each user pain point and related business suggestions. Deploy the large model to the server, interact with other business systems through the API interface, and receive new user comment data for analysis. Continuously collect user feedback, and adjust and optimize the model based on the feedback, such as retraining the model, adjusting hyperparameters, etc. Through the above specific implementation plan, it is possible to accurately analyze user pain points from user comments and generate targeted business suggestions, providing strong support for the company's product improvements and business decisions. The business suggestion report is designed with a title, table of contents, chapter division and other structures to ensure the logic and readability of the report. It is divided into the introduction, market trend analysis, user pain point analysis, and integration results and business suggestions. The introduction introduces the purpose and background of the report, the market trend analysis chapter elaborates on the analysis results of market trend information, the user pain point analysis chapter shows the statistics and analysis of user pain points, and the integration results and business suggestions chapter presents the conclusions after information integration and the business suggestions for these conclusions. In addition, the report is grammatically checked and the layout is optimized to ensure the quality and professionalism of the report. Experts or users in related fields can also be invited to review the report and modify and improve it based on feedback.

[0032] The embodiment of the present application also provides a business suggestion generation system based on a big model, and the business suggestion generation system based on a big model is used in the business suggestion generation method based on a big model. The network information acquired in real time by the business suggestion generation system based on a big model can supplement some of the information missing by business decision makers, so that business decision makers have a more comprehensive understanding of market trends, industry standards, and competitor situations.

[0033] It will be understood by those skilled in the art that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the references to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory ROM, programmable memory PROM, electrically programmable memory DPROM, electrically erasable programmable memory DDPROM or flash memory. Volatile memory may include random access memory RAM or external cache memory.

[0034] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of 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 specification.

[0035] The above embodiments only express several embodiments of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the patent application. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application.

Claims

1. A method for generating business suggestions based on a large model, characterized in that: The method for generating business suggestions based on a large model comprises the following steps: Step S1, obtaining user comment data and business data, and then preprocessing the user comment data to obtain training data; Step S2, building a large model, and then using the training data and the business data to train the large model, to obtain a user pain point analysis model and a business suggestion generation model, wherein the large model includes the user pain point analysis model and the business suggestion generation model; Step S3, inputting the user comment data to be analyzed into the user pain point analysis model to obtain user pain point information; In step S4, the business suggestion generation model obtains market trend information in real time, and then integrates it with user pain point information to generate business suggestions.

2. A method for generating business suggestions based on a large model according to claim 1, characterized in that: In step S1, the method of preprocessing the user comment data to obtain training data includes: Step S11, using OCR text recognition technology combined with API interface technology to obtain user review data from e-commerce platforms, social media and / or product forums; Step S12, using regular expressions and string processing functions to clean the user comment data, and then using the jieba Chinese word segmentation library and / or the nltk English word segmentation library to perform word segmentation processing; Step S13, annotating the data after word segmentation processing based on a pre-established annotation guide, and then dividing the annotated data into a training set, a validation set and a test set according to a preset ratio, so as to obtain training data.

3. The method for generating business suggestions based on a large model according to claim 2, characterized in that: In step S2, a large model is built, and then the large model is trained using training data and business data to obtain a user pain point analysis model and a business suggestion generation model, including: Step S21, select the Transformer architecture and build a large model based on the PyTorch framework; Step S22, defining an input layer, and using a pre-trained word vector model to convert user comment data into a word vector of fixed dimension; Step S23, construct a multi-layer Transformer encoder layer, set the number of heads of the multi-head attention mechanism and the parameters of the feedforward neural network; Step S24, designing the output layer. For the user pain point analysis task, the output layer is designed as a multi-classifier for outputting different types of user pain point information. For the business suggestion generation task, the output layer is designed as a text generation structure for generating relevant business suggestion texts. Step S25, initializing hyperparameters, setting learning rate, batch size, number of training rounds, and hidden layer dimension; Step S26, input the training set and business data into the big model, use the Adam optimizer for training, and obtain the user pain point analysis model and business suggestion generation model.

4. The method for generating business suggestions based on a large model according to claim 3, characterized in that: In step S3, the user comment data to be analyzed is input into the user pain point analysis model, and the method for obtaining the user pain point information includes: Step S31, cleaning and segmenting the user comment data to be analyzed and converting them into word vectors of fixed dimensions, and then inputting them into the user pain point analysis model; In step S32, the user pain point analysis model uses a multi-head attention mechanism to perform semantic association analysis on the words and phrases in the user comment data to be analyzed to understand the user's intention, uses sentiment analysis to determine the sentiment tendency of the user comment data to be analyzed, and uses pattern recognition to discover common patterns and themes in the user comment data to be analyzed to obtain user pain point information.

5. The method for generating business suggestions based on a large model according to claim 4, characterized in that: The step S32 also includes: calculating the cosine similarity between different user pain point information, classifying the user pain point information with similarity exceeding a threshold into one category, and realizing classification and induction of the user pain point information.

6. A method for generating business suggestions based on a large model according to claim 5, characterized in that: In step S4, the method in which the business suggestion generation model obtains market trend information in real time and then integrates it with user pain point information to generate business suggestions includes: Step S41, the business suggestion generation model uses a time series analysis method and a topic model analysis method to obtain and determine the main themes and hot topics in the market trend information; Step S42, counting the occurrence frequencies of different types of user pain point information to determine the priority of the user pain point information; Step S43, comparing the user pain point information with the market trend information one by one according to the priority level and outputting the comparison results, and then integrating the comparison results to generate business suggestions.

7. The method for generating business suggestions based on a large model according to claim 6, characterized in that: In step S43, the method of comparing the user pain point information with the market trend information one by one according to the priority and outputting the comparison result includes: Step S431, comparing the potential connection between the main themes and hot topics in the market trend information and the user pain point information, analyzing whether there is an opportunity to use the market trend information to solve the user pain point and outputting relevant results; Step S432, analyzing whether there is any conflict between the market trend information and the user pain point information and outputting relevant results.

8. The method for generating business suggestions based on a large model according to claim 6, characterized in that: In step S43, the method of integrating the comparison results to generate business suggestions includes: Analyze the correlation between market trend information and user pain point information, then integrate the comparison results between the interrelated market trend information and user pain point information, and comprehensively analyze and elaborate business recommendations.

9. The method for generating business suggestions based on a large model according to claim 8, characterized in that: Methods for integrating the comparison results between interrelated market trend information and user pain point information include: For text information, we use text fusion algorithms to combine and describe interrelated market trend information and user pain point information to generate coherent text paragraphs. For data information, a data fusion algorithm is used to fuse the interrelated market trend information with the user pain point information to generate a new data set.

10. A business suggestion generation system based on a large model, characterized in that: The big model-based business suggestion generation system is used to implement the big model-based business suggestion generation method according to any one of claims 1 to 9.

Images