E-commerce platform commodity image-text content generation method and system based on AI

Abstract

The invention relates to the technical field of image-text generation, and particularly discloses an electronic commerce platform commodity image-text content generation method and system based on AI, and the method comprises the steps: obtaining the multi-source data of an electronic commerce platform commodity, carrying out the feature extraction of the multi-source data based on an AI intelligent algorithm, and obtaining the comprehensive feature information of the commodity, the commodity comprehensive feature information comprises commodity selling point feature information, user preference feature information and market trend feature information, obtaining multiple layers of selling point feature data according to the commodity selling point feature information, and obtaining a commodity selling point evaluation value according to all the layers of selling point feature data. In the data processing link, after multi-source data is obtained, the comprehensive feature information of the commodity is deeply extracted by using the AI intelligent algorithm. Commodity selling points, user preferences and market trend characteristics are accurately extracted from commodity basic information, user evaluation data and market sales volume data, and one-sided cognition caused by insufficient data acquisition and analysis in a traditional mode is avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of image and text generation, and in particular to an AI-based method and system for generating image and text content of products on an e-commerce platform. Background Art

[0002] In the e-commerce industry, product image and text content is the key bridge between consumers and products. It not only carries product information, but is also an important factor in stimulating consumers' desire to buy. However, the traditional way of generating product image and text content faces many difficulties, and emerging technologies based on AI intelligent algorithms are bringing innovative changes to this field.

[0003] The generation of product images and texts on traditional e-commerce platforms mostly relies on manual experience and simple templates. Merchants often select product images based on subjective judgment, and when writing copy, they mainly focus on the basic functions and features of the product, making it difficult to accurately meet the personalized needs of different consumers. Due to the lack of in-depth analysis of consumer preferences, general image and text content cannot attract the attention of specific consumer groups. Taking sports equipment as an example, if the image and text do not highlight the sports scenarios and professional performance suitable for the product, fitness enthusiasts may find it difficult to be interested in it. Summary of the invention

[0004] The purpose of the present invention is to provide an AI-based method and system for generating product image and text content on an e-commerce platform to solve the technical problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions: A method for generating graphic and text content of products on an e-commerce platform based on AI, comprising: Obtain multi-source data of products on e-commerce platforms; Extract features from multi-source data based on AI intelligent algorithms to obtain comprehensive feature information of commodities, wherein the comprehensive feature information of commodities includes feature information of commodity selling points, feature information of user preferences and feature information of market trends; Acquire multiple levels of selling point feature data according to the selling point feature information of the product, and acquire a selling point evaluation rating of the product according to all the selling point feature data of the levels; Acquire multiple preference node feature data according to the user preference feature information, and acquire a user preference fit evaluation value according to each preference node feature data; Acquire a plurality of trend characteristic data according to the market trend characteristic information, and acquire a market trend fit evaluation value according to the plurality of trend characteristic data; The product image and text generation evaluation value is obtained according to the product selling review valuation, user preference fit evaluation value and market trend fit evaluation value, and the adapted e-commerce platform product image and text content is generated according to the product image and text generation evaluation value.

[0006] Preferably, the step of extracting features from multi-source data based on an AI intelligent algorithm to obtain comprehensive feature information of commodities includes: Obtain basic product information, user evaluation data, and market sales data based on multi-source data; Extract product keywords from the basic information of the product based on AI intelligent algorithm; Performing dimensionality reduction screening on the user evaluation data based on the product keywords to obtain screened evaluation information; Acquire user text information and user picture information according to the screening evaluation information; Based on natural language processing technology, text analysis is performed on user text information, and keywords are extracted to obtain text information on potential selling points of products; Identify the user's picture information to obtain key product picture information; Acquire product selling point feature information based on the product potential selling point text information and product key image information; Acquire common information of commodities according to the commodity selling point characteristic information and commodity basic information; Perform sentiment analysis on the common information of the products to obtain users' satisfaction and attention to the products in various aspects; Acquire user preference feature information based on the product satisfaction and product attention; Acquire commodity demand information according to the market sales data, and use the commodity demand information as market trend feature information; The commodity selling point feature information, user preference feature information and market trend feature information are associated and integrated to obtain comprehensive commodity feature information.

[0007] Preferably, the step of obtaining multiple-level selling point feature data according to the product selling point feature information, and obtaining a product selling point review valuation according to all the selling point feature data at all levels includes: Extracting information based on the commodity selling point feature information to obtain a plurality of selling point feature data, wherein the selling point feature data includes function feature data, price feature data, and service feature data; Acquire first characteristic data of the functional characteristic data, and acquire a proportion of the first characteristic data in the first data of the user preference characteristic information; Obtaining second characteristic data of the price characteristic data, and obtaining a proportion of the second characteristic data in the second data of the user preference characteristic information; Obtaining third characteristic data of the service characteristic data, and obtaining a proportion of the third characteristic data in the third data of the user preference characteristic information; The first data proportion value, the second data proportion value and the third data proportion value are obtained to obtain the estimated value of the product sales review.

[0008] Preferably, the step of obtaining a plurality of preference node feature data according to the user preference feature information, and obtaining a user preference fit evaluation value according to each of the preference node feature data comprises: Extracting information according to the user preference characteristics to obtain a plurality of user preference node data; Acquire multiple function bias node data according to the product selling point feature information; Use AI intelligent algorithms to build a user preference model for multiple user preference node data, and input multiple function preference node data into the user preference model; The similarity between the product and the user's preference is obtained according to the user preference model, and the similarity is used as a user preference fit evaluation value.

[0009] Preferably, the step of obtaining a plurality of trend characteristic data according to the market trend characteristic information, and obtaining a market trend fit evaluation value according to the plurality of trend characteristic data comprises: Acquire multiple trend directions according to the market trend characteristic information; Obtain updated adaptability feature data for each trending commodity; Acquire consumption concept fit characteristic data according to the user preference characteristic information; Obtaining trend characteristic data based on the consumption concept fit characteristic data and the product update adaptability characteristic data; Obtaining a trend score according to each of the trend characteristic data; A market trend fit assessment value is obtained based on each trend trend score.

[0010] Preferably, the step of obtaining the product image and text to generate the evaluation value based on the product selling review evaluation value, the user preference fit evaluation value and the market trend fit evaluation value includes: According to the judgment matrix of the analytic hierarchy process for product sales review valuation, user preference fit evaluation value and market trend fit evaluation value, According to the judgment matrix, the first weight value, the second weight value and the third weight value corresponding to the product selling review valuation, the user preference fit evaluation value and the market trend fit evaluation value are obtained; The evaluation value generated by the product picture and text is obtained according to the product selling review valuation, user preference fit evaluation value, market trend fit evaluation value, first weight value, second weight value and third weight value.

[0011] Preferably, the step of generating adapted e-commerce platform product image and text content according to the product image and text generation evaluation value includes: Determine whether the evaluation value generated by the product image and text is within a preset threshold range; If it is, keep the original content of the product image and text; If it is greater than, generating an evaluation value according to the product image and text to obtain a first common feature value; If it is less than, generating an evaluation value according to the product image and text to obtain a second common feature value; Get the preset product image and text content library; Filtering from a preset product image and text content library according to the first common feature value to obtain first image and text update content; Filtering from a preset product image and text content library according to the second common feature value to obtain second image and text update content; Generate product optimization copy and optimized pictures according to the first updated text and picture content or the second updated text and picture content; The optimized copy and optimized pictures of the products are combined with typeset to generate graphic and text content of the products on the e-commerce platform.

[0012] The present invention also provides an electrocardiogram examination operating system for simulation teaching, comprising: The first acquisition module is used to obtain multi-source data of products on the e-commerce platform; The second acquisition module is used to extract features from multi-source data based on an AI intelligent algorithm to obtain comprehensive feature information of commodities, wherein the comprehensive feature information of commodities includes feature information of commodity selling points, feature information of user preferences, and feature information of market trends; A third acquisition module is used to acquire multiple levels of selling point feature data according to the selling point feature information of the product, and acquire a product selling point evaluation value according to all the selling point feature data of the levels; A fourth acquisition module is used to acquire a plurality of preference node feature data according to the user preference feature information, and acquire a user preference fit evaluation value according to each of the preference node feature data; A fifth acquisition module, configured to acquire a plurality of trend characteristic data according to the market trend characteristic information, and acquire a market trend fit evaluation value according to the plurality of trend characteristic data; The generation module is used to obtain a product image and text generation evaluation value based on the product selling review valuation, user preference fit evaluation value and market trend fit evaluation value, and generate adapted e-commerce platform product image and text content based on the product image and text generation evaluation value.

[0013] Preferably, the second acquisition module includes: A first acquisition unit is used to acquire basic product information, user evaluation data and market sales data based on multi-source data; An extraction unit, used to extract commodity keywords from the commodity basic information based on an AI intelligent algorithm; A screening unit, used to perform dimensionality reduction screening on the user evaluation data based on the product keywords to obtain screened evaluation information; A second acquisition unit, used to acquire user text information and user picture information according to the screening evaluation information; A first analysis unit is used to perform text analysis on user text information based on natural language processing technology, and extract keywords to obtain text information of potential selling points of products; An identification unit, used to identify the user's picture information and obtain key picture information of the product; A third acquisition unit is used to acquire commodity selling point feature information according to the commodity potential selling point text information and commodity key image information; A fourth acquisition unit, configured to acquire commodity common information according to the commodity selling point feature information and commodity basic information; A second analysis unit is used to perform sentiment analysis on the common information of the commodities to obtain the user's satisfaction and attention to the commodities in various aspects; A fifth acquisition unit, configured to acquire user preference feature information according to the product satisfaction and product attention; a sixth acquisition unit, configured to acquire commodity demand information according to the market sales data, and use the commodity demand information as market trend feature information; The fusion unit is used to associate and fuse the commodity selling point feature information, user preference feature information and market trend feature information to obtain comprehensive commodity feature information.

[0014] Preferably, the third acquisition module includes: a seventh extraction unit, configured to extract information according to the commodity selling point feature information to obtain a plurality of selling point feature data, wherein the selling point feature data includes function feature data, price feature data, and service feature data; an eighth acquisition unit, configured to acquire first characteristic data of the functional characteristic data, and acquire a proportion of the first characteristic data in the first data of the user preference characteristic information; a ninth acquisition unit, configured to acquire second characteristic data of the price characteristic data, and acquire a proportion of the second characteristic data in the second data of the user preference characteristic information; a tenth acquisition unit, configured to acquire third characteristic data of the service characteristic data, and acquire a proportion of the third characteristic data in the third data of the user preference characteristic information; An eleventh obtaining unit is used to obtain the first data proportion value, the second data proportion value and the third data proportion value to obtain a product selling review valuation.

[0015] The beneficial effects of the present application are as follows: the present invention uses AI intelligent algorithms to deeply extract comprehensive characteristic information of commodities after obtaining multi-source data in the data processing link. The commodity selling points, user preferences and market trend characteristics are accurately extracted from the commodity basic information, user evaluation data and market sales data, avoiding the one-sided cognition caused by insufficient data acquisition and analysis in the traditional way, such as digging out consumers' preferences for smart functions and energy saving and environmental protection from various data of smart home appliances. In the construction of the evaluation system, the commodity selling points, user preference fit and market trend fit are quantitatively evaluated to provide a scientific basis for the generation of commodity graphics and texts. For example, in the field of smart phones, the weights of each evaluation value can be clearly defined, the competitiveness of commodities in the market can be accurately judged, and enterprises can clearly understand the advantages and improvement directions of commodities. In the graphic content generation stage, the graphic content of the commodity on the e-commerce platform that is adapted is generated according to the comprehensive evaluation value. When the evaluation value is in different intervals, suitable materials are selected from the preset commodity graphic content library, and optimized copywriting and pictures are generated and reasonably laid out. This enables graphic and text content to accurately match consumer needs, highlight product advantages, and adapt to market trends. For example, based on the evaluation results of smart wearable devices, the graphics and text can be adjusted to focus on displaying health monitoring functions and fashion designs, thereby effectively attracting target customers and increasing product click-through rates and conversion rates. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 The figure is a schematic diagram of a method flow of an embodiment of the present application.

[0017] Figure 2 A schematic diagram of the system structure of an embodiment of the present application.

[0018] The purpose, features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0019] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0020] like Figure 1 As shown, the present application provides an AI-based method for generating graphic content of goods on an e-commerce platform, which is applied to a circuit component feature database, including: S1. Obtain multi-source data of products on the e-commerce platform; S2. Extracting features from multi-source data based on an AI intelligent algorithm to obtain comprehensive feature information of the product, wherein the comprehensive feature information of the product includes feature information of product selling points, feature information of user preferences, and feature information of market trends; S3, obtaining selling point feature data at multiple levels according to the selling point feature information of the product, and obtaining a selling point evaluation estimate of the product according to all the selling point feature data at the levels; S4, obtaining a plurality of preference node feature data according to the user preference feature information, and obtaining a user preference fit evaluation value according to each of the preference node feature data; S5. Acquire a plurality of trend characteristic data according to the market trend characteristic information, and acquire a market trend fit evaluation value according to the plurality of trend characteristic data; S6. Obtain a product image and text generation evaluation value based on the product sales review valuation, user preference fit evaluation value, and market trend fit evaluation value, and generate an adapted e-commerce platform product image and text content based on the product image and text generation evaluation value.

[0021] As described in the above steps S1-S6, the present invention obtains the multi-source data of the goods on the e-commerce platform through the key step. This step widely collects basic information of goods, user evaluation data, market sales data and market trend data by using API interfaces, crawler technology and cooperating with professional data suppliers. Taking smart watches as an example, not only can basic information such as its brand, function, price, etc. be obtained, but also users' evaluation of its battery life, comfort and technical development trends in the industry can be collected. Then, based on the AI ​​intelligent algorithm, feature extraction is performed on the multi-source data to obtain comprehensive feature information of the goods. This process first cleans and pre-processes the data, and then uses keyword extraction, natural language processing and image recognition technologies to mine potential selling points, user preferences and market trend characteristics of the goods. For example, from the multi-source data of smart watches, the selling points of its health monitoring function, the user's preference for health monitoring functions and the market trend of multi-function long battery life are determined, which overcomes the difficulty of traditional methods in effectively extracting key features, and improves the fit between the graphic content and users and the market. Then, according to the feature information of the selling points of the goods, the selling point feature data at multiple levels are obtained, and the selling point evaluation of the goods is obtained. The selling points of the products are classified by functions, prices, services and other aspects, and the proportion of the characteristic data of each aspect in the user preferences is calculated to obtain the evaluation value. For smart watches, if the functional aspect accounts for a high proportion in the user preferences, it means that its functional advantages are concerned. When creating pictures and texts, this selling point can be highlighted in a targeted manner, solving the problem of incomplete evaluation of traditional selling points. After that, multiple preference node feature data are obtained based on the user preference feature information, and the user preference fit evaluation value is obtained. By determining the preference nodes such as color and function, a user preference model is constructed to calculate the fit between the product and the user preference. For example, for smart watches, if the user prefers simple design and health monitoring functions, and the watch has a high degree of fit with them, the graphic content can be designed around these preferences to increase the user's attention and recognition of the graphic. After that, multiple trend trend feature data are obtained based on the market trend feature information, and the market trend fit evaluation value is obtained. Segment the market trend, obtain data such as product update adaptability and consumer concept fit, and obtain the fit evaluation value through comprehensive scoring. For example, the smartwatch market is popular with blood oxygen monitoring functions and environmentally friendly materials. If the watch has these features, highlighting these trends in the graphic content can enhance its market competitiveness and avoid being out of touch with the market. Finally, the product graphic generation evaluation value is obtained based on the product sales review valuation, user preference fit evaluation value and market trend fit evaluation value, and the adapted e-commerce platform product graphic content is generated. The graphic generation evaluation value is calculated by determining the weight of each evaluation value, and the graphic strategy is determined to generate copy, optimize pictures and layout. For smartwatches, after comprehensive evaluation, the appropriate template is selected to highlight the functions, appearance and market trend related content, improve the quality of graphics and marketing effects, promote product sales, and solve the problem of lack of systematicness in traditional graphic creation.

[0022] In one embodiment, the step of extracting features from multi-source data based on an AI intelligent algorithm to obtain comprehensive feature information of commodities includes: S201, obtaining basic product information, user evaluation data and market sales data based on multi-source data; S202, extracting product keywords from the basic product information based on an AI intelligent algorithm; S203, performing dimensionality reduction screening on the user evaluation data based on the product keywords to obtain screened evaluation information; S204, obtaining user text information and user picture information according to the screening evaluation information; S205, performing text analysis on the user text information based on natural language processing technology, and extracting keywords to obtain text information on potential selling points of the product; S206, identifying the user's picture information to obtain key product picture information; S207, obtaining product selling point feature information according to the product potential selling point text information and product key image information; S208, obtaining commodity common information according to the commodity selling point feature information and commodity basic information; S209, performing sentiment analysis on the common information of the products to obtain the user's satisfaction and attention to the products in various aspects; S210, obtaining user preference feature information according to the product satisfaction and product attention; S211, obtaining commodity demand information according to the market sales data, and using the commodity demand information as market trend feature information; S212: Associating and integrating the commodity selling point feature information, user preference feature information, and market trend feature information to obtain comprehensive commodity feature information.

[0023] As described in the above steps S201-S212, the present invention obtains basic information of goods, user evaluation data and market sales data based on multi-source data. In this step, the basic information such as brand, model, specification, material, etc. of the goods is accurately extracted from the e-commerce platform database through the API interface, and user evaluations are widely collected through multiple channels such as the e-commerce platform comment area, social media, and professional product forums by using crawler technology. At the same time, market sales data is obtained from the backend of the e-commerce platform or professional data statistics agencies. Taking the intelligent sweeping robot as an example, its basic parameters such as cleaning mode and battery capacity are obtained through API; users' evaluations of its cleaning effect and obstacle avoidance ability are collected on various platforms; and sales data of different time periods and regions are obtained from data agencies. Next, the keywords of the goods are extracted from the basic information of the goods based on the AI ​​intelligent algorithm. The AI ​​intelligent algorithm uses natural language processing technology (this technology is a prior art, so it is not described here) to deeply analyze the basic information of the goods and identify words that can represent the core characteristics and advantages of the goods. For example, for an intelligent sweeping robot, the algorithm may extract keywords such as "intelligent route planning", "strong suction", and "automatic recharge". These keywords are like key identifiers, making subsequent processing more targeted and avoiding getting lost in the massive amount of information. The subsequent screening and analysis of user evaluation data can focus on the key features of the product. Then, the user evaluation data is screened based on the product keywords to obtain the screened evaluation information. The extracted product keywords are used as the screening conditions, and the large amount of user evaluation data collected is screened using text matching technology. For the smart sweeping robot, the evaluation content containing keywords such as "intelligent planning route", "strong suction", and "automatic recharge" is screened, and information irrelevant to these key features is removed. In this way, the amount of data is greatly reduced and the analysis efficiency is significantly improved. We can more accurately grasp the user's feedback on the core functions of the product and avoid being disturbed by a large amount of redundant information. After that, the user text information and user picture information are obtained based on the screened evaluation information. Using data classification technology (this technology is an existing technology, so it is not described here), the text and pictures in the screened evaluation information are separated. For smart sweeping robots, user text information may include comments and feedback on product functions, such as "It cleans very well, but gets stuck on carpets" and "The automatic recharging function is very convenient"; user image information may include photos of the sweeping robot working in different scenes, such as cleaning in the living room and automatically recharging. This separation process provides a clear data foundation for further in-depth mining of user feedback and solves the problem that mixed data is difficult to analyze in detail. After that, text analysis of user text information is performed based on natural language processing technology, and keyword extraction is performed to obtain text information on potential selling points of products. With the help of sentiment analysis and keyword extraction functions in natural language processing technology, user text information is deeply mined.Through sentiment analysis, we can determine the sentiment tendency of user reviews, find out the aspects that users are satisfied with, and then extract keywords and phrases from positive reviews as potential selling points. For example, in the user text information of the smart sweeping robot, it is found that many users mentioned that "it can be remotely controlled by mobile phone" and gave positive comments. Then "remote control by mobile phone" can be used as potential selling point text information, which enriches the selling points of the product and makes the product more attractive when it is promoted. Then, the user picture information is identified to obtain the key picture information of the product. The separated user picture information is analyzed using image recognition technology. Through algorithms such as image feature extraction and target detection, picture information that can highlight the characteristics, advantages or usage scenarios of the product is identified. For the smart sweeping robot, pictures showing its flexible obstacle avoidance and cleaning in a complex home environment, pictures of the clean floor after cleaning, etc. may be identified. These pictures, as key picture information, provide a strong reference for optimizing the product picture display and enhance the visual appeal and persuasiveness of the picture and text content. Then, the product selling point feature information is obtained based on the product potential selling point text information and the product key picture information. The potential selling point text information and key image information are associated and integrated to make the selling point of the product more intuitive and comprehensive. For example, the potential selling point of "remote control by mobile phone" is combined with the key image showing the user operating the sweeping robot through the mobile phone to form the selling point feature information of "remote control by mobile phone, control cleaning anytime and anywhere", which can more effectively show the value of the product. Then, the common information of the product is obtained based on the selling point feature information and the basic information of the product. The selling point feature information and the basic information of the product are compared and analyzed to find out the overlap and interrelated content between the two. For the smart sweeping robot, the "laser navigation system" in the basic information of the product corresponds to the "intelligent planning route" in the selling point feature information. "Laser navigation, intelligent planning route" is used as the common information of the product to ensure the consistency of the product at different information levels and improve the overall cognition of the product. Then, the common information of the product is analyzed for sentiment to obtain the user's satisfaction and attention to the product in all aspects. The sentiment analysis technology is used to process the common information of products. By judging the sentiment polarity of user evaluation texts, the user's satisfaction with various aspects of the product can be determined; by counting the frequency of keyword appearance and the number of related evaluations, the user's attention to different aspects of the product can be measured. For example, in the common information of smart sweeping robots, it is found that users have a lot of positive comments on "cleaning effect" and the frequency of mention is high, indicating that users are satisfied with and pay attention to the cleaning effect; while there is attention to "noise control", but there are more negative comments. These analysis results provide a quantitative basis for determining user preferences, so that the graphic content can focus on the key points of user concern. Afterwards, the user preference feature information is obtained based on product satisfaction and product attention.The product satisfaction and product attention data are summarized and analyzed, and the aspects with high user satisfaction and high attention are identified as the main preferences of users, and the aspects with low satisfaction but high attention are identified as the direction that needs to be improved or optimized. For the smart sweeping robot, the analysis results are combined to determine that the user prefers efficient cleaning and low noise operation. When creating graphic content, the advantages of cleaning effect can be highlighted, and reasonable explanations or improvement measures can be proposed for the noise problem to better meet user needs. After that, the product demand information is obtained based on the market sales data, and the product demand information is used as market trend feature information. The market sales data is deeply analyzed to observe the purchase of products in different time periods, different regions, and different consumer groups. Taking the smart sweeping robot as an example, it is found that among young office workers, the sales of products with automatic dust collection function and simple appearance are growing rapidly. From this, the market trend feature information is that young office workers pursue automatic dust collection and simple design. This provides a market orientation for the generation of product graphic content, so that the graphic content can adapt to market changes and enhance the competitiveness of products in the market. Finally, the product selling point feature information, user preference feature information and market trend feature information are associated and integrated to obtain the comprehensive feature information of the product. Use data analysis and information fusion technology to organically integrate these three aspects of information. Analyze the relationship between each piece of information, find out the commonalities and differences, and form comprehensive and systematic product feature information. For the smart sweeping robot, after comprehensive analysis, it is clear that it has selling points such as "laser navigation intelligent route planning, high suction and efficient cleaning, mobile phone remote control, and automatic recharging". Users prefer "efficient cleaning, low noise, automatic dust collection, and simple design". The market trend is that young office workers have an increasing demand for products with relevant functions and designs. This comprehensive feature information provides comprehensive guidance for generating high-quality product graphic content, avoiding highlighting only one aspect in graphic creation and ignoring other important factors, and enhancing the comprehensive competitiveness of products in the market.

[0024] In one embodiment, the step of obtaining multiple levels of selling point feature data according to the product selling point feature information, and obtaining a product selling point review valuation according to all the selling point feature data at each level, includes: S301, extracting information according to the commodity selling point feature information to obtain a plurality of selling point feature data, wherein the selling point feature data includes function feature data, price feature data, and service feature data; S302, obtaining first characteristic data of the functional characteristic data, and obtaining a proportion of the first characteristic data in the first data of the user preference characteristic information; S303, obtaining second characteristic data of the price characteristic data, and obtaining a proportion of the second characteristic data in the second data of the user preference characteristic information; S304, obtaining third characteristic data of the service characteristic data, and obtaining a proportion of the third characteristic data in the third data of the user preference characteristic information; S305: Obtain the first data proportion, the second data proportion and the third data proportion to obtain a product sales review valuation.

[0025] As described in the above steps S301-S305, in the process of obtaining the commodity selling point evaluation evaluation according to the commodity selling point feature information of the present invention, firstly, information extraction is performed according to the commodity selling point feature information to obtain multiple selling point feature data, including functional feature data, price feature data, and service feature data. The commodity selling points are systematically sorted out. For example, for a smart watch, functional feature data is extracted from its many selling points, such as accurate heart rate monitoring and multiple sports mode recognition functions; price feature data, such as its price positioning and cost performance among similar products; service feature data, such as the after-sales maintenance policy provided by the brand, customer service response speed, etc. Through such classification and extraction, the problem of cluttered selling point information is solved, and a clear foundation is laid for subsequent evaluation. Then, the first feature data of the functional feature data is obtained, and the first data proportion value of the first feature data in the user preference feature information is obtained. Taking the smart watch as an example, the accuracy of heart rate monitoring is selected as the first feature data. By analyzing the preference information of a large number of users on the smart watch, the proportion of the number of times the accuracy of heart rate monitoring is mentioned in the total number of preference information is counted to obtain the first data proportion value. This step can clarify the focus of users on product functions, avoid blindly emphasizing functions during publicity and promotion, and allow functional advantages to reach target users more accurately. Then, obtain the second feature data of the price feature data, and obtain the second data proportion of the second feature data in the user preference feature information. For smart watches, use its cost performance as the second feature data, and calculate the proportion of cost performance related content in the total preference information by studying the user's preference information on price and performance. Doing so will help companies accurately grasp the user's sensitivity to product prices, avoid price strategies deviating from user needs, highlight price advantages in promotion, and enhance the price competitiveness of products. After that, obtain the third feature data of the service feature data, and obtain the third data proportion of the third feature data in the user preference feature information. For example, the two-year warranty service provided by smart watch brands is used as the third feature data. By collecting the user's preference information on after-sales service expectations and evaluations, the frequency of mentioning the two-year warranty is counted as the proportion of the total preference information frequency. This can help companies understand users' concerns and importance of services, optimize service quality, highlight service advantages in publicity, and enhance users' purchasing confidence. Finally, the first data proportion value, the second data proportion value, and the third data proportion value are obtained to obtain the product sales review valuation. Combining these three proportion values ​​is like giving a comprehensive "scoring" of the attractiveness of the product in terms of function, price, and service. For example, by comprehensively considering the proportion of smart watches in user preferences in terms of heart rate monitoring accuracy, cost-effectiveness, and warranty services, an evaluation value reflecting the attractiveness of its overall selling points is obtained.

[0026] In one embodiment, the step of obtaining a plurality of preference node feature data according to the user preference feature information, and obtaining a user preference fit evaluation value according to each of the preference node feature data, comprises: S401, extracting information according to the user preference characteristics to obtain a plurality of user preference node data; S402, acquiring multiple function bias node data according to the product selling point feature information; S403, using AI intelligent algorithm to build a user preference model for multiple user preference node data, and inputting multiple function preference node data into the user preference model; S404: Obtain the similarity between the product and the user preference according to the user preference model, and use the similarity as a user preference fit evaluation value.

[0027] As described in the above steps S401-S404, in the process of obtaining the user preference fit evaluation value according to the user preference feature information, the present invention first extracts information according to the user preference feature to obtain multiple user preference node data. This step is crucial. By using technologies such as data mining and natural language processing, specific preference information can be extracted from a large amount of data such as user purchase records, evaluation feedback, and search keywords. For example, in the field of smart home products, it is found through analysis that users often mention the requirements of "high security" and "sensitive fingerprint recognition" for smart door locks and "adjustable color temperature" for smart lamps. These have become preference node data such as "high security preference", "sensitive fingerprint recognition preference", and "adjustable color temperature preference". This effectively solves the problem that user preference information is disorganized and difficult to use directly, and lays the foundation for the subsequent accurate grasp of user needs. Then, multiple function bias node data are obtained according to the product selling point feature information. For various types of products, their selling points are numerous and complicated. This step is to screen out the key functional advantages and convert them into function bias node data. Taking smart speakers as an example, they may have selling points such as excellent sound quality, intelligent voice interaction, and rich content resources. After in-depth analysis, the "high-fidelity sound quality function bias node", "accurate voice recognition function bias node", "massive music resource function bias node", etc. are determined. This makes the core functions of the product stand out, avoids the dilemma of information redundancy in marketing and product positioning, allows companies to more clearly understand the value of their own products, and provides a clear reference for matching with user preferences. Then, the AI ​​intelligent algorithm is used to build a user preference model for multiple user preference node data, and multiple function bias node data are input into the user preference model. The AI ​​intelligent algorithm can deeply analyze the internal connections and rules between user preference node data, so as to build a model that fits the real needs of users. Taking online education products as an example, after collecting user preference node data such as "strong interactivity", "complete curriculum system", and "strong teaching staff", the AI ​​intelligent algorithm is used to build a model. Then, the "real-time interactive classroom function bias node", "professional course arrangement function bias node", "senior teacher teaching function bias node" of a certain online education product are input into the model. This process solves the problem that traditional methods are difficult to comprehensively and accurately evaluate the fit between products and user preferences, allowing companies to analyze the match between products and user needs from a scientific and quantitative perspective. Finally, the similarity between products and user preferences is obtained based on the user preference model, and the similarity is used as the user preference fit evaluation value. Through the analysis of the user preference model, the functional preference node data of the product and the user preference node data are compared to intuitively derive the similarity between the two, that is, the user preference fit evaluation value. For example, in the beauty product market, if the user prefers "natural ingredients" and "long-lasting makeup effect", the selling point of a certain foundation is "plant essence formula" and "lasting makeup all day long", and the model evaluation will give a higher fit evaluation value.This provides key quantitative basis for enterprises in product promotion, improvement and market positioning. Enterprises can accurately recommend products to target users based on the evaluation value, optimize product design to better meet user needs, avoid blind decision-making, and enhance market competitiveness.

[0028] In one embodiment, the step of obtaining a plurality of trend characteristic data according to the market trend characteristic information, and obtaining a market trend fit evaluation value according to the plurality of trend characteristic data includes: S501, obtaining multiple trend directions according to the market trend characteristic information; S502, obtaining update adaptability characteristic data of commodities of each trend direction; S503, acquiring consumption concept compatibility characteristic data according to the user preference characteristic information; S504, acquiring trend characteristic data according to the consumption concept fit characteristic data and the commodity update adaptability characteristic data; S505, obtaining a trend score according to each of the trend characteristic data; S506: Obtain a market trend fit evaluation value according to each trend trend score.

[0029] As described in the above steps S501-S506, in the process of obtaining the market trend fit evaluation value according to the market trend feature information, the present invention first obtains multiple trend directions according to the market trend feature information. With the help of data analysis technology, enterprises widely collect industry reports, market research data, policy and regulatory dynamics, and technology development information and other information, and sort out various trends such as the popularization of 5G technology in the smartphone market, which has led to an increase in demand for 5G mobile phones, consumers' requirements for mobile phone camera functions have increased, and foldable screen mobile phones have gradually emerged. This allows companies to clearly see the direction of market development and avoid getting lost in complex market information. It provides forward-looking guidance for corporate strategic planning and product development, and then obtains the product update adaptability feature data for each trend. The company conducts in-depth analysis of each identified trend in combination with its own product characteristics. For example, in the field of smart wearable devices, facing the trend of "enhanced health monitoring functions", companies compare the gap between the existing health monitoring functions of their own smart bracelets (only heart rate and step monitoring) and market demand (adding blood pressure monitoring, deep sleep analysis and other functions), so as to obtain the updated adaptability characteristic data of the product under this trend, clarify the direction of product improvement, and avoid blindly investing resources in product updates. Then, according to the user preference characteristic information, obtain the consumer concept fit characteristic data. Companies use market research and data analysis methods to collect user feedback through questionnaires, user evaluation analysis, social media monitoring and other channels, and extract changes in consumer consumption concepts and preferences. Take the cosmetics market as an example. It is found that consumers are increasingly concerned about ingredient safety and personalized customization. A cosmetics brand obtains the fit characteristic data with consumer needs in these aspects based on this, providing precise directions for product development and market promotion, and preventing products and marketing from being out of touch with consumer needs. After that, according to the consumer concept fit characteristic data and product update adaptability characteristic data, obtain trend trend characteristic data. The company conducts correlation analysis on these two types of data and compares the differences and connections between the two. Taking the automobile market as an example, we comprehensively consider consumers' demand for new energy and intelligence, as well as the current status of a certain automobile brand in new energy technology and intelligent configuration, find out the points of convergence and gap between products and market demand, and obtain more targeted trend characteristic data such as "new energy technology needs to accelerate research and development, and intelligent configuration needs to be greatly improved" to ensure that corporate decisions are more scientific and reasonable and can respond quickly to market changes. After that, we obtain trend scores based on each trend characteristic data. Enterprises establish a scientific and reasonable scoring system to score key indicators of trend characteristic data from multiple dimensions such as market demand, the degree of matching between products and trends, and the competitive advantages of enterprises in this trend.In the notebook computer market, for the trend of "lightness and high performance", a brand of notebook computers is scored after comprehensive consideration of product lightness, performance, heat dissipation technology, etc., so that companies can intuitively understand the performance of products in various trends, identify advantages and disadvantages, and provide a clear direction for resource allocation. Finally, the market trend fit evaluation value is obtained based on the score of each trend. According to the scores of each trend, the company considers the importance of different trends and uses appropriate methods (such as assigning weights to each trend according to market research and corporate strategy) to calculate the market trend fit evaluation value. In the smart home market, a brand comprehensively considers the scores and weights of trends such as "intelligent upgrade", "energy saving and environmental protection", and "personalized customization" to obtain a market trend fit evaluation value. This quantitative indicator can comprehensively evaluate the overall fit between products and market trends.

[0030] In one embodiment, the step of obtaining the product image and text to generate the evaluation value based on the product selling review evaluation value, the user preference fit evaluation value and the market trend fit evaluation value includes: S601, judging the matrix of product selling review valuation, user preference fit evaluation value and market trend fit evaluation value according to the hierarchical analysis method, S602, obtaining a first weight value, a second weight value, and a third weight value corresponding to the product selling review valuation, the user preference fit evaluation value, and the market trend fit evaluation value according to the judgment matrix; S603, obtaining a product image and text generation evaluation value based on the product selling review valuation, user preference fit evaluation value, market trend fit evaluation value, first weight value, second weight value and third weight value.

[0031] As described in the above steps S601-S603, in the process of generating evaluation values ​​by acquiring product images and texts, the present invention first constructs a judgment matrix based on the analytic hierarchy process for the product selling review valuation, user preference fit evaluation value and market trend fit evaluation value. This step is crucial. It compares the three key factors that affect the generation of product images and texts in pairs through comprehensive expert experience, market research and historical data to determine their relative importance. For example, in the field of smart home appliances, after in-depth analysis, it was found that under the current market environment, consumers' functional requirements for smart home appliances change rapidly, so market trend fit is relatively important; and satisfying users' preferences for convenient operation and personalized functions is also critical; although the selling point of the product itself is important, it is slightly inferior to the first two. Based on these judgments, the market trend fit is compared with the product selling points and user preference fit respectively to determine their importance relationship, and then a judgment matrix is ​​constructed. This allows the company to clearly understand the relative position of each factor, avoid one-sided view of a certain factor in subsequent decision-making, and lay the foundation for scientific decision-making. Then, according to the constructed judgment matrix, the weight values ​​corresponding to the product selling review valuation, user preference fit evaluation value and market trend fit evaluation value are obtained. Using a specific calculation method, the judgment matrix is ​​processed, and the weight corresponding to each evaluation value is obtained while ensuring the rationality and reliability of the judgment matrix. For example, after calculation, the weight corresponding to the product selling review valuation in the field of smart home appliances may be 0.2, the weight corresponding to the user preference fit evaluation value is 0.3, and the weight corresponding to the market trend fit evaluation value is 0.5. These weights clarify the proportion of each factor in the evaluation of product image and text generation, so that when enterprises create images and texts, they can reasonably allocate resources according to the weights, highlight key factors, avoid waste of resources, and improve the pertinence and attractiveness of images and texts. Finally, the product image and text generation evaluation value is obtained by combining the product selling review valuation, user preference fit evaluation value, market trend fit evaluation value and their corresponding weight values. In this step, each evaluation value is multiplied by its weight and then added to obtain an evaluation value that can fully reflect the comprehensive performance of the product image and text. Taking a certain smart sweeping robot as an example, if its product selling point evaluation value is 7 points, the user preference fit evaluation value is 8 points, and the market trend fit evaluation value is 6 points, combined with the weights mentioned above, the product image and text generation evaluation value can be calculated. This evaluation value provides an intuitive and quantitative indicator for enterprises, which can be used to judge the quality and effect of current images and texts. If the evaluation value is high, it means that the graphics and text perform well in highlighting selling points, meeting user preferences and keeping up with market trends; if the evaluation value is low, the company can adjust the graphics and text content in a targeted manner, such as supplementing new function introductions based on market trends, optimizing expressions related to user preferences, etc., so as to improve the quality of graphics and text and promote product sales.

[0032] In one embodiment, the step of generating adapted e-commerce platform product image and text content according to the product image and text generation evaluation value includes: S604, determining whether the evaluation value generated by the product image and text is within a preset threshold range; If it is, keep the original content of the product image and text; If it is greater than, generating an evaluation value according to the product image and text to obtain a first common feature value; If it is less than, generating an evaluation value according to the product image and text to obtain a second common feature value; S605, obtaining a preset product image and text content library; S606: Filter from a preset product image and text content library according to the first common feature value to obtain first image and text update content; S607: Filter from a preset product image and text content library according to the second common feature value to obtain second image and text update content; S608: Generate a product optimization copy and an optimized picture according to the first updated text and picture content or the second updated text and picture content; S609: Combining the optimized product copy and the optimized image to generate product image and text content on the e-commerce platform.

[0033] As described in the above steps S604-S609, the present invention generates an adapted e-commerce platform commodity graphic content based on the commodity graphic generation evaluation value. The first step is to determine whether the commodity graphic generation evaluation value is within the preset threshold range. The enterprise sets an evaluation value range based on its own business conditions, market feedback on the quality of commodity graphics, and past data and other comprehensive factors. After that, the actual calculated commodity graphic generation evaluation value is compared with it. For example, an e-commerce company selling digital products sets a threshold range of 7-9 points. If the evaluation value of a headset is 8 points, which is within the range, it means that the current headset graphic quality meets the standard and can be maintained as it is; if a tablet computer has an evaluation value of 9.5 points, which is greater than the upper limit, it is necessary to explore its advantages to obtain the first common feature value; if a smart bracelet has an evaluation value of only 6 points, which is lower than the lower limit, it is necessary to find the problem and obtain the second common feature value. This step clarifies for the enterprise whether the graphics need to be optimized and the direction of optimization, avoiding blind adjustments, and then obtains the preset commodity graphic content library. In daily operations, companies continue to accumulate graphic materials for various products, such as product detail pictures, usage scenario pictures, function description copy, promotional copy, etc. After that, they are classified and sorted according to dimensions such as product type, style, and applicable scenarios to build a preset product graphic content library. This is like building a material warehouse. When it is necessary to optimize product graphics, companies can quickly find suitable materials from it, effectively solving the problem of insufficient materials and improving optimization efficiency. Then, according to the first common feature value, the preset product graphic content library is screened to obtain the first graphic update content. When the evaluation value is greater than the upper limit of the threshold interval, the advantages of the graphics are analyzed in depth to determine the first common feature value. For example, a smart watch has a high evaluation value. After analysis, it is found that its advantages lie in its accurate health monitoring function and fashionable appearance design. This is the first common feature value. Based on this, matching materials are screened in the content library, such as texts that describe the health monitoring function in detail, pictures that show the fashionable appearance from multiple angles, etc., to form the first updated text and picture content, further highlighting the advantages of the product text and pictures, and enhancing its attractiveness and competitiveness. After that, the second common feature value is selected from the preset product text and picture content library to obtain the second updated text and picture content. When the evaluation value is less than the lower limit of the threshold interval, the problems existing in the text and pictures are carefully analyzed to determine the second common feature value. For example, a certain Bluetooth speaker has a low evaluation value. After inspection, it is found that the text description of the sound quality effect is vague, and the picture does not show the unique design of the speaker. This is the second common feature value. Subsequently, materials that can solve these problems are searched in the content library, such as texts that describe the sound quality characteristics in detail, pictures that show the unique design of the speaker, etc., to form the second updated text and picture content, to optimize the text and pictures in a targeted manner, and to make up for the shortcomings. After that, the product optimized text and optimized pictures are generated according to the first text and picture update content or the second text and picture update content.If the content is updated based on the first picture and text, the materials that highlight the advantages are integrated into the original copy, and the copy structure and expression are adjusted to make it more vivid and vivid to show the advantages of the product. At the same time, pictures that can highlight the advantages are selected and processed; if the content is updated based on the second picture and text, the copy is modified and supplemented according to the problems of the picture and text, and the relevant pictures are replaced or optimized. For example, the smart watch updates the content based on the first picture and text, elaborates on the principles and advantages of the health monitoring function in the copy, and optimizes the pictures showing the appearance; the Bluetooth speaker updates the content based on the second picture and text, improves the sound quality description copy, replaces the pictures of the display design, and allows the pictures to convey the value of the product more effectively. Finally, the optimized product copy and the optimized pictures are combined with typesetting to generate the e-commerce platform product picture and text content. Enterprises choose the appropriate typesetting method based on the display rules of the e-commerce platform, such as picture size, word limit of the copy, etc., such as picture and text, picture-dominated or copy-dominated. When typesetting, pay attention to the coordination between the copy and the picture, reasonably arrange the position, font, color and other elements, and use blank space to enhance the visual effect. For example, when displaying a mobile phone, a layout with both pictures and texts is used, with high-definition pictures of the phone from different angles displayed at the top and corresponding texts introducing each function in detail at the bottom, ensuring that the generated picture and text content complies with platform specifications.

[0034] like Figure 2 As shown, the present invention also provides a simulated teaching electrocardiogram examination operating system, comprising: The first acquisition module 1 is used to acquire multi-source data of products on the e-commerce platform; The second acquisition module 2 is used to extract features from multi-source data based on an AI intelligent algorithm to obtain comprehensive feature information of commodities, wherein the comprehensive feature information of commodities includes feature information of commodity selling points, feature information of user preferences, and feature information of market trends; A third acquisition module 3 is used to acquire multiple levels of selling point feature data according to the product selling point feature information, and acquire a product selling point evaluation value according to all the selling point feature data at the levels; A fourth acquisition module 4 is used to acquire a plurality of preference node feature data according to the user preference feature information, and acquire a user preference fit evaluation value according to each of the preference node feature data; A fifth acquisition module 5, configured to acquire a plurality of trend characteristic data according to the market trend characteristic information, and acquire a market trend fit evaluation value according to the plurality of trend characteristic data; The generation module 6 is used to obtain the product image and text generation evaluation value based on the product selling review valuation, user preference fit evaluation value and market trend fit evaluation value, and generate adapted e-commerce platform product image and text content based on the product image and text generation evaluation value.

[0035] In one embodiment, the second acquisition module 2 includes: A first acquisition unit is used to acquire basic product information, user evaluation data and market sales data based on multi-source data; An extraction unit, used to extract commodity keywords from the commodity basic information based on an AI intelligent algorithm; A screening unit, used to perform dimensionality reduction screening on the user evaluation data based on the product keywords to obtain screened evaluation information; A second acquisition unit, used to acquire user text information and user picture information according to the screening evaluation information; A first analysis unit is used to perform text analysis on user text information based on natural language processing technology, and extract keywords to obtain text information of potential selling points of products; An identification unit, used to identify the user's picture information and obtain key picture information of the product; A third acquisition unit is used to acquire commodity selling point feature information according to the commodity potential selling point text information and commodity key image information; A fourth acquisition unit, configured to acquire commodity common information according to the commodity selling point feature information and commodity basic information; A second analysis unit is used to perform sentiment analysis on the common information of the commodities to obtain the user's satisfaction and attention to the commodities in various aspects; A fifth acquisition unit, configured to acquire user preference feature information according to the product satisfaction and product attention; a sixth acquisition unit, configured to acquire commodity demand information according to the market sales data, and use the commodity demand information as market trend feature information; The fusion unit is used to associate and fuse the commodity selling point feature information, user preference feature information and market trend feature information to obtain comprehensive commodity feature information.

[0036] In one embodiment, the third acquisition module 3 includes: a seventh extraction unit, configured to extract information according to the commodity selling point feature information to obtain a plurality of selling point feature data, wherein the selling point feature data includes function feature data, price feature data, and service feature data; an eighth acquisition unit, configured to acquire first characteristic data of the functional characteristic data, and acquire a proportion of the first characteristic data in the first data of the user preference characteristic information; a ninth acquisition unit, configured to acquire second characteristic data of the price characteristic data, and acquire a proportion of the second characteristic data in the second data of the user preference characteristic information; a tenth acquisition unit, configured to acquire third characteristic data of the service characteristic data, and acquire a proportion of the third characteristic data in the third data of the user preference characteristic information; An eleventh obtaining unit is used to obtain the first data proportion value, the second data proportion value and the third data proportion value to obtain a product selling review valuation.

[0037] Those of ordinary skill in the art can understand 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, any reference to memory, storage, value library or other media provided in this application and used in the embodiments may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0038] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, device, article or method including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, device, article or method. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the presence of other identical elements in the process, device, article or method including the element.

[0039] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. Any equivalent results or equivalent process changes made using the contents of the present invention description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for generating product graphic content on an e-commerce platform based on AI, characterized in that: include: Obtain multi-source data of products on e-commerce platforms; Extracting features from multi-source data based on AI intelligent algorithms to obtain comprehensive feature information of commodities, wherein the comprehensive feature information of commodities includes feature information of commodity selling points, feature information of user preferences and feature information of market trends; Acquire multiple levels of selling point feature data according to the selling point feature information of the product, and acquire a selling point evaluation rating of the product according to all the selling point feature data of the levels; Acquire multiple preference node feature data according to the user preference feature information, and acquire a user preference fit evaluation value according to each preference node feature data; Acquire a plurality of trend characteristic data according to the market trend characteristic information, and acquire a market trend fit evaluation value according to the plurality of trend characteristic data; The product image and text generation evaluation value is obtained according to the product selling review valuation, user preference fit evaluation value and market trend fit evaluation value, and the adapted e-commerce platform product image and text content is generated according to the product image and text generation evaluation value.

2. According to claim 1, a method for generating product graphic content on an e-commerce platform based on AI is characterized in that: The step of extracting features from multi-source data based on an AI intelligent algorithm to obtain comprehensive feature information of commodities includes: Obtain basic product information, user evaluation data, and market sales data based on multi-source data; Extract product keywords from the basic information of the product based on AI intelligent algorithm; Performing dimensionality reduction screening on the user evaluation data based on the product keywords to obtain screened evaluation information; Acquire user text information and user picture information according to the screening evaluation information; Based on natural language processing technology, text analysis is performed on user text information, and keywords are extracted to obtain text information on potential selling points of products; Identify the user's picture information to obtain key product picture information; Acquire product selling point feature information based on the product potential selling point text information and product key image information; Acquire common information of commodities according to the commodity selling point characteristic information and commodity basic information; Perform sentiment analysis on the common information of the products to obtain users' satisfaction and attention to the products in various aspects; Acquire user preference feature information based on the product satisfaction and product attention; Acquire commodity demand information according to the market sales data, and use the commodity demand information as market trend feature information; The commodity selling point feature information, user preference feature information and market trend feature information are associated and integrated to obtain comprehensive commodity feature information.

3. The method for generating product graphic content on an e-commerce platform based on AI according to claim 1, characterized in that: The step of obtaining multiple levels of selling point feature data according to the product selling point feature information, and obtaining a product selling point review valuation according to all the selling point feature data at all levels, includes: Extracting information based on the commodity selling point feature information to obtain a plurality of selling point feature data, wherein the selling point feature data includes function feature data, price feature data, and service feature data; Acquire first characteristic data of the functional characteristic data, and acquire a proportion of the first characteristic data in the first data of the user preference characteristic information; Obtaining second characteristic data of the price characteristic data, and obtaining a proportion of the second characteristic data in the second data of the user preference characteristic information; Obtaining third characteristic data of the service characteristic data, and obtaining a proportion of the third characteristic data in the third data of the user preference characteristic information; The first data proportion value, the second data proportion value and the third data proportion value are obtained to obtain the estimated value of the product sales review.

4. The method for generating product graphic content on an e-commerce platform based on AI according to claim 1, characterized in that: The step of acquiring a plurality of preference node feature data according to the user preference feature information, and acquiring a user preference fit evaluation value according to each of the preference node feature data, comprises: Extracting information according to the user preference characteristics to obtain a plurality of user preference node data; Acquire multiple function bias node data according to the product selling point feature information; Use AI intelligent algorithms to build a user preference model for multiple user preference node data, and input multiple function preference node data into the user preference model; The similarity between the product and the user's preference is obtained according to the user preference model, and the similarity is used as a user preference fit evaluation value.

5. The method for generating product graphic content on an e-commerce platform based on AI according to claim 1, characterized in that: The step of obtaining a plurality of trend characteristic data according to the market trend characteristic information, and obtaining a market trend fit evaluation value according to the plurality of trend characteristic data comprises: Acquire multiple trend directions according to the market trend characteristic information; Obtain updated adaptability feature data for each trending commodity; Acquire consumption concept fit characteristic data according to the user preference characteristic information; Obtaining trend characteristic data based on the consumption concept fit characteristic data and the product update adaptability characteristic data; Obtaining a trend score according to each of the trend characteristic data; A market trend fit assessment value is obtained based on each trend trend score.

6. The method for generating product graphic content on an e-commerce platform based on AI according to claim 1, characterized in that: The step of obtaining the product image and text to generate the evaluation value based on the product selling review evaluation value, the user preference fit evaluation value and the market trend fit evaluation value comprises: According to the judgment matrix of the analytic hierarchy process for product sales review valuation, user preference fit evaluation value and market trend fit evaluation value, According to the judgment matrix, the first weight value, the second weight value and the third weight value corresponding to the product selling review valuation, the user preference fit evaluation value and the market trend fit evaluation value are obtained; The evaluation value generated by the product picture and text is obtained according to the product selling review valuation, user preference fit evaluation value, market trend fit evaluation value, first weight value, second weight value and third weight value.

7. The method for generating product graphic content on an e-commerce platform based on AI according to claim 1, characterized in that: The step of generating adapted e-commerce platform product image and text content according to the product image and text generation evaluation value includes: Determine whether the evaluation value generated by the product image and text is within a preset threshold range; If it is, keep the original content of the product image and text; If it is greater than, generating an evaluation value according to the product image and text to obtain a first common feature value; If it is less than, generating an evaluation value according to the product image and text to obtain a second common feature value; Get the preset product image and text content library; Filtering from a preset product image and text content library according to the first common feature value to obtain first image and text update content; Filtering from a preset product image and text content library according to the second common feature value to obtain second image and text update content; Generate product optimization copy and optimized pictures according to the first updated text and picture content or the second updated text and picture content; The optimized copy and optimized pictures of the products are combined with typeset to generate graphic and text content of the products on the e-commerce platform.

8. A simulated teaching electrocardiogram examination operating system, characterized in that: include: The first acquisition module is used to obtain multi-source data of products on the e-commerce platform; The second acquisition module is used to extract features from multi-source data based on an AI intelligent algorithm to obtain comprehensive feature information of commodities, wherein the comprehensive feature information of commodities includes feature information of commodity selling points, feature information of user preferences, and feature information of market trends; A third acquisition module is used to acquire multiple levels of selling point feature data according to the selling point feature information of the product, and acquire a product selling point evaluation value according to all the selling point feature data of the levels; A fourth acquisition module is used to acquire a plurality of preference node feature data according to the user preference feature information, and acquire a user preference fit evaluation value according to each of the preference node feature data; A fifth acquisition module, configured to acquire a plurality of trend characteristic data according to the market trend characteristic information, and acquire a market trend fit evaluation value according to the plurality of trend characteristic data; The generation module is used to obtain a product image and text generation evaluation value based on the product selling review valuation, user preference fit evaluation value and market trend fit evaluation value, and generate adapted e-commerce platform product image and text content based on the product image and text generation evaluation value.

9. The electrocardiogram examination operating system for simulation teaching according to claim 7, characterized in that: The second acquisition module includes: A first acquisition unit is used to acquire basic product information, user evaluation data and market sales data based on multi-source data; An extraction unit, used to extract commodity keywords from the commodity basic information based on an AI intelligent algorithm; A screening unit, used to perform dimensionality reduction screening on the user evaluation data based on the product keywords to obtain screened evaluation information; A second acquisition unit, used to acquire user text information and user picture information according to the screening evaluation information; A first analysis unit is used to perform text analysis on user text information based on natural language processing technology, and extract keywords to obtain text information of potential selling points of products; An identification unit, used to identify the user's picture information and obtain key picture information of the product; A third acquisition unit is used to acquire commodity selling point feature information according to the commodity potential selling point text information and commodity key image information; A fourth acquisition unit, configured to acquire commodity common information according to the commodity selling point feature information and commodity basic information; A second analysis unit is used to perform sentiment analysis on the common information of the commodities to obtain the user's satisfaction and attention to the commodities in various aspects; A fifth acquisition unit, configured to acquire user preference feature information according to the product satisfaction and product attention; a sixth acquisition unit, configured to acquire commodity demand information according to the market sales data, and use the commodity demand information as market trend feature information; The fusion unit is used to associate and fuse the commodity selling point feature information, user preference feature information and market trend feature information to obtain comprehensive commodity feature information.

10. The electrocardiogram examination operating system for simulation teaching according to claim 7, characterized in that: The third acquisition module includes: a seventh extraction unit, configured to extract information according to the commodity selling point feature information to obtain a plurality of selling point feature data, wherein the selling point feature data includes function feature data, price feature data, and service feature data; an eighth acquisition unit, configured to acquire first characteristic data of the functional characteristic data, and acquire a proportion of the first characteristic data in the first data of the user preference characteristic information; a ninth acquisition unit, configured to acquire second characteristic data of the price characteristic data, and acquire a proportion of the second characteristic data in the second data of the user preference characteristic information; a tenth acquisition unit, configured to acquire third characteristic data of the service characteristic data, and acquire a proportion of the third characteristic data in the third data of the user preference characteristic information; An eleventh obtaining unit is used to obtain the first data proportion value, the second data proportion value and the third data proportion value to obtain a product selling review valuation.

Images