Commodity recommendation method and device, storage medium and electronic device

Abstract

The embodiment of the invention provides a commodity recommendation method and device, a storage medium and an electronic device.The method comprises the steps that obtained social media data are analyzed to obtain commodity data of commodities included in the social media data, and the social media data are data screened out based on target keywords; determining a target commodity matched with commodity information included in the commodity data from a target commodity library; acquiring historical sales data of the target commodity; determining a business reference index of the target commodity based on the historical sales data and the commodity data; and displaying the service reference index on the target interface to indicate the target object to determine a recommended commodity from the target commodities through the service reference index. Through the method and the device, the problem of inaccurate e-commerce live broadcast commodity selection is solved, and the effect of accurately selecting e-commerce live broadcast commodities is further achieved.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of computers, and more particularly, to a method, apparatus, storage medium, and electronic device for recommending products. Background Art

[0002] In the related art, the product selection tools in the e-commerce live streaming industry mainly rely on historical data analysis and market trend analysis, and cannot reflect the current consumer market in a timely manner. In addition, their capabilities in processing complex data and generating personalized recommendations are significantly insufficient.

[0003] It can be seen that there is a technical problem of inaccurate product selection in e-commerce live streaming in the related art.

[0004] At present, there is no effective solution to the above problems existing in the related art. Summary of the Invention

[0005] The embodiments of the present invention provide a method, apparatus, storage medium, and electronic device for recommending products, so as to at least solve the problem of inaccurate product selection in e-commerce live streaming existing in the related art.

[0006] According to an embodiment of the present invention, there is provided a method for recommending products, including: parsing the obtained social media data to obtain product data of products included in the social media data, where the social media data is data filtered based on a target keyword; determining a target product matching the product information included in the product data from a target product library; obtaining historical sales data of the target product; determining a business reference index of the target product based on the historical sales data and the product data; and displaying the business reference index on a target interface to instruct a target object to determine a recommended product from the target products based on the business reference index.

[0007] In an exemplary embodiment, determining the business reference index of the target product based on the historical sales data and the product data includes: batch-processing the historical sales data to determine historical sales information of the target product, where the historical sales information includes at least one of the following: the sales quantity of the target product, the gross profit margin, the number of covered stores, and the number of covered stores is used to indicate the number of stores selling the target product; stream-processing the product data to determine real-time data of the target product, where the real-time data includes at least one of the following: product popularity, product subscription information, and product purchase tendency; and determining the historical sales information and the real-time data as the business reference index.

[0008] In an exemplary embodiment, when the real-time data includes the popularity of the product, stream processing is performed on the product data to determine that the real-time data of the target product includes: determining the number of likes, comments, collections, and shares of the target product, where the number of likes represents the number of times the target product is liked, the number of comments represents the number of times the target product is commented on, the number of collections represents the number of times the target product is collected, and the number of shares represents the number of times the target product is shared; determining the product popularity included in the real-time data based on the number of likes, the number of comments, the number of collections, and the number of shares.

[0009] In an exemplary embodiment, determining the product popularity included in the real-time data based on the number of likes, the number of comments, the number of collections, and the number of shares includes: determining a first weight corresponding to the number of likes; determining a second weight corresponding to the number of comments; determining a third weight corresponding to the number of collections; determining a fourth weight corresponding to the number of shares; determining a first product of the number of likes and the first weight; determining a second product of the number of comments and the second weight; determining a third product of the number of collections and the third weight; determining a fourth product of the number of shares and the fourth weight; and determining the sum value of the first product, the second product, the third product, and the fourth product as the product popularity.

[0010] In an exemplary embodiment, parsing the obtained social media data to obtain the product data of the products included in the social media data includes: generating a product prompt for extracting the product data, where the product prompt is a prompt generated based on the chain-of-thought pattern; and parsing out the product data from the social media data according to the product prompt.

[0011] In an exemplary embodiment, generating a product prompt for extracting the product data includes: determining whether the title of the social media data includes a product keyword related to a preset keyword; in the case of including the product keyword, determining the target content included in the body of the social media data for describing the product keyword; summarizing the target content to obtain a summary content; determining the list of promoted products included in the social media data; determining the product categories of the promoted products included in the list of promoted products; determining the positive content and the negative content of the promoted products included in the social media data; determining the recommendation intention for the promoted products included in the social media data; determining the price information of the promoted products included in the social media data; and generating a piece of product data in a target format for each promoted product, where the product data includes the summary content, the product category, the positive content, the negative content, the recommendation intention, and the price information.

[0012] In an exemplary embodiment, after the business reference metrics are displayed on the target interface to indicate that the target object determines a recommended product from the target product based on the business reference metrics, the method further includes: when the positive content and the negative content of the recommended product are non-empty in the product data, generating a recommended reason in a target format based on the positive content and the negative content according to the target language style; when the positive content and the negative content of the recommended product are empty in the product data, generating a recommended reason in a target format according to the target language style.

[0013] According to another embodiment of the present invention, there is provided a product recommendation device, including: a parsing module, configured to parse the obtained social media data to obtain product data of products included in the social media data, where the social media data is data filtered based on a target keyword; a first determination module, configured to determine a target product matching the product information included in the product data from a target product library; an obtaining module, configured to obtain historical sales data of the target product; a second determination module, configured to determine business reference metrics of the target product based on the historical sales data and the product data; a display module, configured to display the business reference metrics on a target interface to indicate that a target object determines a recommended product from the target product based on the business reference metrics.

[0014] According to still another embodiment of the present invention, there is further provided a computer-readable storage medium, in which a computer program is stored, where the computer program is configured to execute the steps in any one of the above method embodiments when running.

[0015] According to still another embodiment of the present invention, there is further provided an electronic device, including a memory and a processor, where a computer program is stored in the memory, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0016] According to still another embodiment of the present invention, there is further provided a computer program product, including a computer program, where the steps of the methods described in various embodiments of the present application are implemented when the computer program is executed by a processor.

[0017] Through the present invention, the social media data obtained by screening keywords can be parsed first to obtain the product data of the products included in the social media data. The target product that matches the product information included in the product data is determined from the target product library, and the historical sales data of this target product is obtained. The business reference indicators of this target product can be determined through the historical sales data and the product data, and this business reference indicator is displayed on the target interface, that is, the target object can be instructed to determine the recommended product from the target products according to this business reference indicator. Since the social media data can timely reflect the user's demand for different target products, and the recommended products are selected in combination with the business reference indicators, the product selection strategy can be dynamically adjusted according to the market demand. Therefore, the problem of inaccurate product selection in e-commerce live broadcasts in the related art can be solved, and the effect of accurately selecting e-commerce live broadcast products can be achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 is a block diagram of the hardware structure of a mobile terminal for a product recommendation method according to an embodiment of the present invention;

[0019] Figure 2 is a flowchart of a product recommendation method according to an embodiment of the present invention;

[0020] Figure 3 is a flowchart of social media data parsing according to an embodiment of the present invention;

[0021] Figure 4 is a flowchart of product data mapping according to an embodiment of the present invention;

[0022] Figure 5 is a flowchart of product mapping result process processing according to an embodiment of the present invention;

[0023] Figure 6 is an effect diagram of a product selection dashboard according to an embodiment of the present invention;

[0024] Figure 7 is a flowchart of big data service processing according to an embodiment of the present invention;

[0025] Figure 8 is an effect diagram of product recommendation reasons according to an embodiment of the present invention;

[0026] Figure 9 is a block diagram of the structure of a product recommendation device according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0027] In the following, embodiments of the present invention will be described in detail with reference to the drawings and in conjunction with the embodiments.

[0028] It should be noted that the terms "first", "second", etc. in the description, claims and the above-mentioned drawings of the present invention are used to distinguish similar objects, and do not necessarily have to be used to describe a specific order or sequence.

[0029] The method embodiments provided in the embodiments of the present application can be executed on a mobile terminal, a computer terminal or a similar computing device. Taking the operation on a mobile terminal as an example, Figure 1 is a hardware structure block diagram of a mobile terminal for a product recommendation method according to an embodiment of the present invention. As Figure 1 shown, the mobile terminal may include one or more ( Figure 1 only one is shown in the figure) processors 102 (the processor 102 may include, but is not limited to, a processing device such as a microprocessor MCU or a programmable logic device FPGA) and a memory 104 for storing data. Among them, the above-mentioned mobile terminal may further include a transmission device 106 for communication functions and an input / output device 108. Those of ordinary skill in the art can understand that Figure 1 the structure shown is only schematic and does not limit the structure of the above-mentioned mobile terminal. For example, the mobile terminal may further include more or fewer components than those shown in Figure 1 the figure, or have a different configuration from that shown in Figure 1 the figure.

[0030] The memory 104 can be used to store computer programs. For example, software programs and modules of application software, such as the computer program corresponding to the product recommendation method in the embodiments of the present invention. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, that is, the above-mentioned method is implemented. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memories, or other non-volatile solid-state memories. In some instances, the memory 104 may further include a memory remotely provided with respect to the processor 102, and these remote memories can be connected to the mobile terminal through a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an enterprise internal network, a local area network, a mobile communication network, and combinations thereof.

[0031] The transmission device 106 is used to receive or send data via a network. Specific examples of the above-mentioned network may include a wireless network provided by a communication provider of the mobile terminal. In one instance, the transmission device 106 includes a network adapter (abbreviated as NIC), which can be connected to other network devices through a base station and thus can communicate with the Internet. In one instance, the transmission device 106 may be a radio frequency (abbreviated as RF) module, which is used to communicate with the Internet wirelessly.

[0032] In this embodiment, a method for recommending products is provided. Figure 2 It is a flowchart of the method for recommending products according to the embodiment of the present invention, as Figure 2 shown. The process includes the following steps:

[0033] Step S202: Analyze the obtained social media data to obtain the product data of the products included in the social media data, where the social media data is data filtered based on a target keyword;

[0034] Step S204: Determine target products that match the product information included in the product data from a target product library;

[0035] Step S206: Obtain the historical sales data of the target products;

[0036] Step S208: Determine the business reference indicators of the target products based on the historical sales data and the product data;

[0037] Step S210: Display the business reference indicators on a target interface to instruct a target object to determine recommended products from the target products based on the business reference indicators.

[0038] In the above embodiment, social media data can be obtained through web crawler technology or in the form of cooperation with a third-party company. Social media data can be understood as social media data related to a specified subscription term (i.e., the above target keyword). Among them, web crawler technology can be understood as an automatic program used to automatically access, build indexes for search engines, and extract information on the Internet, and crawl web page content according to specific rules. Social media can include video software, picture software, or chat software with product sales attributes. The target keyword can be casual snacks, baked goods, mother and baby good products, or clothing items, etc. Figure 3 It is a flowchart of social media data analysis according to the embodiment of the present invention, as Figure 3 shown. After the obtained social media data is processed such as filtering, deduplication, and deleting empty data, it can be analyzed using an AI (Artificial Intelligence) large model. Product data, brand data, etc. included in the social media data can be obtained, and sentiment analysis can also be performed on the social media data to extract data such as the advantages and disadvantages of products and purchase tendencies. Among them, the AI large model can be understood as an artificial intelligence model with a large number of parameters and a complex structure, which can process and learn complex data patterns based on deep learning technology.

[0039] In the above embodiment, the AI large model can continuously analyze social media data and store the parsed product data, brand data, etc. in real time. Through the mapping of product data, effective target products can be obtained.Figure 4 is a flowchart of commodity data mapping according to an embodiment of the present invention. As Figure 4 shown, the AI large model can perform data mapping with the commodity library of a specified merchant or store (i.e., the above-mentioned target commodity library), parse the commodity name, and return the commodity. For the schematic diagram of the commodity return result, please refer to Figure 5 , Figure 5 is a flowchart of the process for processing the commodity mapping result according to an embodiment of the present invention. As Figure 5 shown, the commodities with high commodity matching degree and high interaction data shown in the commodity mapping result can be used as the target commodities for product selection recommendations, which can be used to flexibly provide commodity recommendations for live broadcast product selection; for the commodities with low commodity matching degree but high interaction data, they can be used as the direction for subsequent new product mining, which can be used to continuously expand and optimize the merchant's commodity categories.

[0040] In the above embodiment, the historical commodity sales data of the existing target commodities can be obtained according to different business scenarios, and multiple business reference indicators of the target commodities can be determined based on the historical commodity sales data and the commodity data. After the big data platform completes the processing of the historical commodity sales data and the real-time data, a product selection dashboard (i.e., the above-mentioned target interface) of the target commodities can be built on the BI (Business Intelligence) platform. Figure 6 is an effect diagram of the product selection dashboard according to an embodiment of the present invention. As Figure 6 shown, Figure 6 (a) It can display the heat ranking data of different target keywords (i.e., the above-mentioned business reference indicators) in real time. Figure 6 (b) It can display the number of comments corresponding to different brands in real time. The target object before the e-commerce live broadcast can complete the preliminary product selection work according to the business reference indicators displayed on this target interface, that is, recommended commodities can be selected. For sudden hot commodities, commodities with a relatively high matching degree can be promptly put on the shelves, enhancing the market response speed and being able to quickly capture the changes in consumer demands. Among them, the BI platform can be understood as a collection of software and technical tools, which can help enterprises collect, process, analyze, and visualize data, so as to convert massive data into easily understandable information to support decision-making.

[0041] Through the present invention, the social media data obtained by screening keywords can be parsed first to obtain the product data of the products included in the social media data. The target product that matches the product information included in the product data is determined from the target product library, and the historical sales data of this target product is obtained. The business reference index of this target product can be determined through the historical sales data and the product data, and this business reference index is displayed on the target interface, that is, the target object can be instructed to determine the recommended products from the target products according to this business reference index. Since the social media data can timely reflect the user's demand for different target products, and the recommended products are selected in combination with the business reference index, the product selection strategy can be dynamically adjusted according to the market demand. Therefore, the problem of inaccurate product selection in e-commerce live broadcasts in the related art can be solved, and the effect of accurately selecting e-commerce live broadcast products can be achieved.

[0042] Optionally, the execution subject of the above steps may be a terminal, or a server, a client, etc., but not limited thereto.

[0043] In an exemplary embodiment, determining the business reference index of the target product based on the historical sales data and the product data includes: batch-processing the historical sales data to determine the historical sales information of the target product, where the historical sales information includes at least one of the following: the sales volume of the target product, the gross profit margin, the number of covered stores, and the number of covered stores is used to indicate the number of stores selling the target product; stream-processing the product data to determine the real-time data of the target product, where the real-time data includes at least one of the following: product popularity, product subscription information, and product purchase tendency; determining the historical sales information and the real-time data as the business reference index.

[0044] In the above embodiment, the AI large model can parse product information, interaction data, product advantages and disadvantages, purchase tendency, mapped products, etc., and synchronize the data to the big data platform through real-time big data tasks. In the big data platform, quasi-real-time business processing can be achieved by batch-processing historical sales data and stream-processing product data. Figure 7 It is a big data business processing flow chart according to an embodiment of the present invention, as Figure 7As shown, the big data service can batch process historical sales data according to different business scenarios, and obtain the historical sales information of the target product: the sales data of the target product, the gross profit margin, the number of covered stores online and the proportion. By counting the sales volume, the market acceptance and sales potential of the target product can be evaluated; by calculating the gross profit margin, its profitability can be understood; by the covered store data and the proportion, the market penetration rate of the target product can be reflected indirectly. Among them, batch processing can be understood as a method for processing a large amount of data, which can be used for one-time processing of large-scale data sets. Its core feature is to divide the data into multiple smaller data blocks and then batch process these data blocks, which can improve efficiency and resource utilization. Stream processing can be understood as a technology for real-time processing and analysis of rapidly arriving data streams, aiming to quickly obtain valuable information. Its main features can include high speed, high throughput and low latency, and it is suitable for application scenarios that require real-time response.

[0045] In the above embodiment, the AI large model can also process product data, that is, perform stream processing on the product data, and obtain the real-time data of the target product: product popularity, product subscription information, and product purchase tendency. By calculating the product popularity, the current popularity of the product can be reflected; by the product subscription information, the subscription situation of consumers for specific products or categories can be analyzed, and potential interests and consumption demands can be understood; by the product purchase tendency, the recommendation tendency of consumers for products can be identified.

[0046] In the above embodiment, integrating the historical sales information obtained by batch processing and the real-time data obtained by stream processing can form business reference indicators, which can provide a comprehensive market analysis for e-commerce live broadcasts, help merchants make more accurate product selection and marketing decisions before and during the live broadcast, and thus improve the efficiency and revenue of the live broadcast.

[0047] In an exemplary embodiment, when the real-time data includes the product popularity, performing stream processing on the product data to determine that the real-time data of the target product includes: determining the number of likes, comments, collections, and shares of the target product, where the number of likes represents the number of times the target product is liked, the number of comments represents the number of times the target product is commented, the number of collections represents the number of times the target product is collected, and the number of shares represents the number of times the target product is shared; determining the product popularity included in the real-time data based on the number of likes, the number of comments, the number of collections, and the number of shares.

[0048] In the above embodiments, the popularity of a product, as a core indicator for product selection recommendations and new product discovery, may include cumulative popularity and incremental popularity. Among them, the incremental popularity can be used to show the upward trend of a product's popularity in a short period of time, which can facilitate real-time adjustment of the live product selection strategy. Therefore, calculating the product popularity value in real time can timely reflect consumers' consumption tendencies.

[0049] In the above embodiments, streaming processing of product data can obtain interaction data, such as the number of likes, comments, collections, and shares of a target product. Among them, the number of likes represents the number of times the target product has been liked, the number of comments represents the number of times the target product has been commented on, the number of collections represents the number of times the target product has been collected, and the number of shares represents the number of times the target product has been shared. Through the number of likes, comments, collections, and shares, real-time data can be determined, that is, by real-time monitoring and calculating the popularity value of the target product, accurate market insights can be provided for merchants, and more effective product recommendation decisions can be made during the live broadcast.

[0050] In an exemplary embodiment, determining the product popularity included in the real-time data based on the number of likes, the number of comments, the number of collections, and the number of shares includes: determining a first weight corresponding to the number of likes; determining a second weight corresponding to the number of comments; determining a third weight corresponding to the number of collections; determining a fourth weight corresponding to the number of shares; determining a first product of the number of likes and the first weight; determining a second product of the number of comments and the second weight; determining a third product of the number of collections and the third weight; determining a fourth product of the number of shares and the fourth weight; and determining the sum value of the first product, the second product, the third product, and the fourth product as the product popularity.

[0051] In the above embodiments, the calculation of product popularity can be adjusted in real time according to different business scenarios and different data indicators. Taking the calculation of product popularity based on simple interaction data as a reference: First, according to business requirements and market analysis, the weight values corresponding to the number of likes, comments, collections, and shares can be determined. Different weight values reflect the contribution values of different interaction types to product popularity, and the weights can be determined according to the importance of the interaction types of the target product. For example, the number of comments and the number of shares can reflect deeper user participation and dissemination willingness, so higher weights can be assigned. That is, the second weight can be 2, and the fourth weight can be 3; the number of likes and the number of collections reflect users' initial interest in the target product, and the weights are relatively low. That is, the first weight can be 1, and the third weight can be 1.5. Then the calculation formula for product popularity can be expressed as: number of likes × 1 + number of comments × 2 + number of collections × 1.5 + number of shares × 3, where number of likes × 1 is the above-mentioned first product; number of comments × 2 is the above-mentioned second product; number of collections × 1.5 is the above-mentioned third product; number of shares × 3 is the above-mentioned fourth product.

[0052] In an exemplary embodiment, parsing the obtained social media data to obtain the product data of the products included in the social media data includes: generating a product prompt for extracting the product data, where the product prompt is a prompt generated based on the chain-of-thought pattern; parsing the product data from the social media data according to the product prompt.

[0053] In the above embodiment, CoT (Chain of Thought) can be used to generate a product prompt for interacting with the AI large model, which can ensure that the model can gradually reason according to the product prompt and accurately extract information such as products in the social media data, enabling the AI large model to efficiently extract valuable information from complex texts and providing effective support for product selection in e-commerce live broadcasts. Among them, a prompt can be understood as a specific instruction or question input by a user in the field of artificial intelligence, which can be used to guide the model to generate corresponding outputs. By providing clear context and instructions, the AI system can more accurately understand the user's intention and make a response. The product prompt can contain a series of instructions to guide the model to identify and extract relevant information from the social media data; the CoT chain-of-thought pattern can be understood as an artificial intelligence technology that can simulate the thinking process of humans when solving problems. By gradually constructing a logical chain from the problem to the answer, it emphasizes the transparency and interpretability of the reasoning process, making the decision of the AI closer to the human thinking mode.

[0054] In an exemplary embodiment, generating a product prompt for extracting the product data includes: determining whether the title of the social media data includes a product keyword related to a pre-set keyword; in the case of including the product keyword, determining the target content included in the body of the social media data for describing the product keyword; summarizing the target content to obtain a summary content; determining the list of promoted products included in the social media data; determining the product category of the promoted products included in the list of promoted products; determining the positive content and negative content of the promoted products included in the social media data; determining the recommendation intention for the promoted products included in the social media data; determining the price information of the promoted products included in the social media data; generating a piece of product data in a target format for each promoted product, where the product data includes the summary content, the product category, the positive content, the negative content, the recommendation intention, and the price information.

[0055] In the above embodiment, the template of the product prompt is as follows:

[0056]

[0057]

[0058]

[0059]

[0060] In the above embodiments, the specific description of the main steps of the product prompt words is as follows:

[0061] (1) Definition of the role: The system will be defined as the name "Content Analysis Expert", and the role responsibility "Product Information Extraction Expert" is clearly defined. With a professional tone and medium friendliness, it is good at identifying products, brands, classifications, and promotion intentions from text content.

[0062] (2) Structured output: Through step-by-step instructions, each step clearly defines the specific tasks to be completed, from checking keywords to extracting price information, which can ensure clear logic and enable the model to process information step by step.

[0063] (3) Step-by-step reasoning: The subsequent instructions are based on the analysis results of the previous instructions. For example, after analyzing the title, content analysis is carried out, which can form a coherent thinking chain and make the final output more accurate.

[0064] (4) Transparency and feedback: By clearly defining the output requirements for each step (such as return format and character limit), the transparency of the model output results can be enhanced, and the model is allowed to perform self-verification based on intermediate results.

[0065] (5) Standardized format: By using JSON format for output, the information of each product exists independently, making the data easy to parse and use.

[0066] In this embodiment, first, a set of pre-set keywords subscribe_word can be determined according to business requirements, such as bread, baking, cake, and biscuits, etc. The AI model will detect whether the product keywords associated with the pre-set keywords are included in the title titile of the social media data. In the case of including such relevant product keywords, the title title and the body content content can be further analyzed to determine the target content describing the product keywords associated with the keywords subscribe_word, such as the characteristics of the product, usage scenarios, brand information, etc. Summarize the extracted target content to obtain a summary content of no more than 100 characters and return the summary content.

[0067] In this embodiment, it is possible to comprehensively analyze and summarize whether the content mainly promotes product keywords, and return the list of promoted products in the summary content, ensuring that each product keyword is included in the promoted product list. If a certain product is not promoted, the promoted product list is empty. Subsequently, it is possible to analyze the specific categories of the promoted products included in the promoted product list (i.e., the above-mentioned product categories), and return the category list. It is also possible to determine the positive reviews (i.e., the above-mentioned positive content) and negative reviews (i.e., the above-mentioned negative content) of the promoted products included in the social media data, extract the advantages pros of the positive reviews and the disadvantages cons of the negative reviews, and perform deduplication.

[0068] In this embodiment, it is also possible to determine whether to recommend purchasing this promoted product, and return the recommendation intention: the value is "recommend", "do not recommend", or "unknown"; extract the price information price of the promoted product in the social media data and return it. For each promoted product, an independent JSON target format record (i.e., the above-mentioned product data) is formed, which may include: brand, summary content, product category, positive evaluation content and advantages, negative evaluation content and disadvantages, summary, recommendation intention, and price information. Among them, the JSON format can be understood as a lightweight data exchange format, which is easy for humans to read and write, and is also convenient for machines to parse and generate. Its basic structure consists of key-value pairs.

[0069] In an exemplary embodiment, after displaying the business reference indicators on the target interface to instruct the target object to determine the recommended products from the target products based on the business reference indicators, the method further includes: when the positive content and negative content of the recommended products are non-empty in the product data, generating recommended reasons in a target format based on the positive content and the negative content according to the target language style; when the positive content and negative content of the recommended products are empty in the product data, generating recommended reasons in a target format according to the target language style.

[0070] In the above embodiment, the AI large model can also generate recommended reasons for the recommended products to assist the target object in product marketing. Still, the thought chain mode is adopted to generate the prompt words for interacting with the AI large model. The template for generating the prompt words for product recommended reasons is as follows:

[0071]

[0072]

[0073] In the above embodiment, the specific description of the main steps of the template for generating the prompt words for product recommended reasons is as follows:

[0074] (1) Task Definition: The task of the system is to generate marketing copy based on the input product name, positive reviews, and negative reviews. The generated content will be returned in the recommended reasons to assist merchants in effective product promotion. The system is defined as name "Product Explanation Specialist" with a clear role of "Shennan Engineering Marketing Copy", using a professional tone and medium friendliness, and being proficient in nutrition knowledge.

[0075] (2) Content Analysis: The system first analyzes the input positive and negative reviews, extracts the key information such as the satisfaction points and recommended reasons of consumers for the product, and generates marketing copy.

[0076] (3) Application of Nutrition Knowledge: If both the input positive and negative reviews are empty, the system will utilize its nutrition knowledge to generate appropriate marketing copy based on the product name. This process ensures that even without user feedback, the system can still provide valuable information to attract potential consumers.

[0077] (4) Output Format: The finally generated marketing copy will be output in standard JSON format to ensure a clear data structure and easy parsing.

[0078] In this embodiment, the defined system task is to generate marketing copy based on the input product name goodsName, positive content positive_reviews, and negative content negative_reviews. First, if the input positive content and negative content are not empty, the system can analyze the input positive content and negative content, extract the key information to generate recommended reasons; if the input positive content and negative content are empty, it can use nutrition knowledge to generate appropriate recommended reasons corresponding to the product name. In addition, during the process of generating recommended reasons, in order to attract consumers, the copy can be output in a kind and humorous language style (i.e., the above target language style) in the standard JSON format (i.e., the above target format), which can efficiently generate marketing copy that meets market needs and consumer psychology and provide strong sales support for merchants. Among them, the effect diagram of the recommended reasons for the product can be referred to Figure 8 , such as Figure 8 shown, and the recommended reasons can include the category name of each recommended product, the tendency of consumers, and the recommended reasons corresponding to each product.

[0079] In the above embodiments, by using an AI large model to parse social media data in real time, information such as the popularity of products and the preferences of consumers can be quickly obtained; by integrating multiple data sources (such as historical sales records, etc.), comprehensive market insights can be provided to help merchants make more informed decisions; based on the data analysis results, product recommendations that meet market demands and consumer psychology can be dynamically generated, which can improve the accuracy and efficiency of product selection and facilitate timely adjustment of strategies. In addition, automatically generating recommendation reasons that match the product features can enhance the interaction between the target object and consumers and improve the conversion rate of sales.

[0080] Through the description of the above implementation manners, those skilled in the art can clearly understand that the method according to the above embodiments can be implemented by means of software plus a necessary general hardware platform. Of course, it can also be implemented by hardware, but in many cases, the former is a better implementation manner. Based on such an understanding, the technical solution of the present invention, in essence, or the part that makes contributions to the prior art, can be embodied in the form of a software product. The computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes several instructions for causing a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in various embodiments of the present invention.

[0081] In this embodiment, a product recommendation device is also provided. This device is used to implement the above embodiments and preferred implementation manners, and those that have been described will not be repeated. As used hereinafter, the term "module" can be a combination of software and / or hardware that can achieve a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementation in hardware, or a combination of software and hardware is also possible and contemplated.

[0082] Figure 9 is a structural block diagram of a product recommendation device according to an embodiment of the present invention. As Figure 9 shown, the device includes:

[0083] A parsing module 902, configured to parse the obtained social media data to obtain product data of products included in the social media data, where the social media data is data screened based on a target keyword;

[0084] A first determination module 904, configured to determine a target product that matches the product information included in the product data from a target product library;

[0085] An acquisition module 906, configured to acquire historical sales data of the target product;

[0086] A second determination module 908, configured to determine a business reference index of the target product based on the historical sales data and the product data;

[0087] A display module 910 displays the business reference metrics on a target interface to indicate that a target object determines a recommended product from the target product based on the business reference metrics.

[0088] In an exemplary embodiment, the second determination module 908 may implement determining the business reference metrics of the target product based on the historical sales data and the product data in the following manner: performing batch processing on the historical sales data to determine the historical sales information of the target product, where the historical sales information includes at least one of the following: the sales quantity of the target product, the gross profit margin, the number of covered stores, and the number of covered stores is used to indicate the number of stores selling the target product; performing stream processing on the product data to determine the real-time data of the target product, where the real-time data includes at least one of the following: product popularity, product subscription information, and product purchase tendency; and determining the historical sales information and the real-time data as the business reference metrics.

[0089] In an exemplary embodiment, when the real-time data includes the product popularity, the second determination module 908 may implement performing stream processing on the product data to determine the real-time data of the target product in the following manner: determining the number of likes, the number of comments, the number of collections, and the number of shares of the target product, where the number of likes represents the number of times the target product is liked, the number of comments represents the number of times the target product is commented on, the number of collections represents the number of times the target product is collected, and the number of shares represents the number of times the target product is shared; and determining the product popularity included in the real-time data based on the number of likes, the number of comments, the number of collections, and the number of shares.

[0090] In an exemplary embodiment, the second determination module 908 may implement determining the product popularity included in the real-time data based on the number of likes, the number of comments, the number of collections, and the number of shares in the following manner: determining a first weight corresponding to the number of likes; determining a second weight corresponding to the number of comments; determining a third weight corresponding to the number of collections; determining a fourth weight corresponding to the number of shares; determining a first product of the number of likes and the first weight; determining a second product of the number of comments and the second weight; determining a third product of the number of collections and the third weight; determining a fourth product of the number of shares and the fourth weight; and determining the sum value of the first product, the second product, the third product, and the fourth product as the product popularity.

[0091] In an exemplary embodiment, the parsing module 902 may parse the obtained social media data to obtain the product data of the products included in the social media data in the following manner: generate product prompting words for extracting the product data, where the product prompting words are prompting words generated based on the chain of thought pattern; and parse the product data from the social media data according to the product prompting words.

[0092] In an exemplary embodiment, the parsing module 902 may generate product prompting words for extracting the product data in the following manner: determine whether the title of the social media data includes product keywords related to pre-set keywords; in the case of including the product keywords, determine the target content included in the body of the social media data for describing the product keywords; summarize the target content to obtain a summary content; determine the list of promoted products included in the social media data; determine the product categories of the promoted products included in the list of promoted products; determine the positive content and the negative content of the promoted products included in the social media data; determine the recommendation intention for the promoted products included in the social media data; determine the price information of the promoted products included in the social media data; and generate a piece of product data in a target format for each promoted product, where the product data includes the summary content, the product category, the positive content, the negative content, the recommendation intention, and the price information.

[0093] In an exemplary embodiment, the apparatus may be used to display the business reference metrics on a target interface to indicate that after a target object determines a recommended product from the target products based on the business reference metrics: when the positive content and the negative content of the recommended product are non-empty in the product data, generate a recommendation reason in a target language style based on the positive content and the negative content; and when the positive content and the negative content of the recommended product are empty in the product data, generate a recommendation reason in a target language style.

[0094] It should be noted that the above-mentioned respective modules may be implemented by software or hardware. For the latter, it may be implemented in the following manner, but not limited thereto: all the above-mentioned modules are located in the same processor; or, the above-mentioned respective modules are located in different processors in any combination form.

[0095] An embodiment of the present invention further provides a computer-readable storage medium, in which a computer program is stored, where the computer program is configured to execute the steps in any one of the above-mentioned method embodiments when running.

[0096] In an exemplary embodiment, the above computer-readable storage medium may include, but is not limited to: various media that can store computer programs such as USB flash drives, read-only memory (ROM for short), random access memory (RAM for short), mobile hard disks, magnetic disks, or optical discs.

[0097] An embodiment of the present invention also provides an electronic device, including a memory and a processor. A computer program is stored in the memory, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0098] In an exemplary embodiment, the above electronic device may further include a transmission device and input / output devices. Among them, the transmission device is connected to the above processor, and the input / output devices are connected to the above processor.

[0099] An embodiment of the present invention also provides a computer program product, including a computer program. When the computer program is executed by a processor, it implements the steps of the methods in various embodiments of the present application.

[0100] The specific examples in this embodiment may refer to the examples described in the above embodiments and exemplary embodiments, and will not be repeated here.

[0101] Obviously, those skilled in the art should understand that the above modules or steps of the present invention can be implemented by a general-purpose computing device. They can be concentrated on a single computing device or distributed on a network composed of multiple computing devices. They can be implemented by program codes executable by the computing device. Thus, they can be stored in a storage device and executed by the computing device. And in some cases, the steps shown or described can be executed in a different order than here, or they can be separately made into individual integrated circuit modules, or multiple modules or steps among them can be made into a single integrated circuit module to implement. In this way, the present invention is not limited to any specific combination of hardware and software.

[0102] The above are only the preferred embodiments of the present invention and are not used to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for recommending a product, characterized in that: include: Parsing the acquired social media data to obtain commodity data of commodities included in the social media data, wherein the social media data is data screened based on target keywords; Determine, from the target commodity library, a target commodity that matches the commodity information included in the commodity data; Obtaining historical sales data of the target product; Determine the business reference index of the target commodity based on the historical sales data and the commodity data; The business reference indicator is displayed on the target interface to instruct the target object to determine the recommended product from the target products through the business reference indicator.

2. The method according to claim 1, characterized in that: Determining the business reference index of the target commodity based on the historical sales data and the commodity data includes: Batch processing is performed on the historical sales data to determine the historical sales information of the target product, wherein the historical sales information includes at least one of the following: the sales volume, gross profit margin, and number of covered stores of the target product, wherein the number of covered stores is used to indicate the number of stores that sell the target product; Performing stream processing on the commodity data to determine real-time data of the target commodity, wherein the real-time data includes at least one of the following: commodity popularity, commodity subscription information, and commodity purchase tendency; The historical sales information and the real-time data are determined as the business reference indicators.

3. The method according to claim 2, characterized in that In the case where the real-time data includes the popularity of the product, stream processing is performed on the product data to determine that the real-time data of the target product includes: Determine the number of likes, comments, favorites, and shares of the target product, wherein the number of likes indicates the number of times the target product has been liked, the number of comments indicates the number of times the target product has been commented on, the number of favorites indicates the number of times the target product has been favorited, and the number of shares indicates the number of times the target product has been shared; The popularity of the product included in the real-time data is determined based on the number of likes, the number of comments, the number of favorites, and the number of shares.

4. The method according to claim 3, characterized in that Determining the popularity of the product included in the real-time data based on the number of likes, the number of comments, the number of favorites, and the number of shares includes: Determine a first weight corresponding to the number of likes; Determine a second weight corresponding to the number of comments; Determine a third weight corresponding to the number of favorites; Determine a fourth weight corresponding to the number of shares; Determine a first product of the number of likes and the first weight; Determine a second product of the number of comments and the second weight; Determine a third product of the number of favorites and the third weight; determining a fourth product of the number of shares and the fourth weight; The sum of the first product, the second product, the third product and the fourth product is determined as the commodity popularity.

5. The method according to claim 1, characterized in that Parsing the acquired social media data to obtain commodity data of commodities included in the social media data includes: Generating a commodity prompt word for extracting the commodity data, wherein the commodity prompt word is a prompt word generated based on a thought chain model; The product data is parsed from the social media data according to the product prompt word.

6. The method according to claim 5, characterized in that Generating a product prompt word for extracting the product data includes: Determining whether the title of the social media data includes a product keyword related to a preset keyword; In a case where the product keyword is included, determining target content included in the body of the social media data and used to describe the product keyword; Summarizing the target content to obtain summary content; Determining a list of promoted products included in the social media data; Determining the product categories of the promotional products included in the promotional product list; Determining positive content and negative content of the promoted product included in the social media data; Determining the recommendation intention for the promoted product included in the social media data; Determining price information of the promoted product included in the social media data; A piece of product data in a target format is generated for each of the promoted products, wherein the product data includes the summary content, the product category, the positive content, the negative content, the recommendation intention, and the price information.

7. The method according to claim 1, characterized in that After displaying the business reference indicator on the target interface to instruct the target object to determine the recommended product from the target products according to the business reference indicator, the method further includes: When the positive content and the negative content of the recommended product included in the product data are not empty, generating a recommendation reason in a target format based on the positive content and the negative content according to the target language style; When the positive content and the negative content of the recommended product included in the product data are empty, a recommendation reason in a target format is generated according to a target language style.

8. A device for recommending products, characterized in that: include: A parsing module, used to parse the acquired social media data to obtain commodity data of commodities included in the social media data, wherein the social media data is data screened based on target keywords; A first determination module, configured to determine a target commodity matching the commodity information included in the commodity data from a target commodity library; An acquisition module, used to acquire the historical sales data of the target product; A second determination module determines a business reference indicator of the target commodity based on the historical sales data and the commodity data; The display module displays the business reference indicator on the target interface to instruct the target object to determine the recommended product from the target products through the business reference indicator.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, wherein the computer program is configured to execute the method according to any one of claims 1 to 7 when executed.

10. An electronic device comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to run the computer program to perform the method according to any one of claims 1 to 7.

11. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method described in any one of claims 1 to 7 are implemented.

Images