Product recommendation method and apparatus, electronic device, and storage medium

Abstract

The application relates to the technical field of big data, and discloses a product recommendation method and device, electronic equipment and a storage medium, which comprise the following steps: calculating a first similarity value of vector representation between a first label of a first product and a second label of a second product; if the first similarity value is greater than a first preset threshold value, determining a user list to be recommended according to order data of the second product; obtaining behavior data of each user to be recommended in the user list to be recommended; judging whether the behavior data contains access data of a third product related to first attribute information; if yes, constructing a third label of the third product according to the access data; calculating a second similarity value of vector representation between the first label and the third label; obtaining a target recommended user from the user list to be recommended according to the second similarity value; and recommending the first product to the target recommended user. The application can ensure that a new product can be accurately recommended to a user in the case that order data of the new product is lacking.

Description

Technical Field

[0001] This invention relates to the field of big data technology, and in particular to a product recommendation method, apparatus, electronic device, and storage medium. Background Technology

[0002] In the fintech sector, companies frequently launch new products, new packages, or new benefits packages to meet the ever-changing needs of users.

[0003] Existing recommendation methods rely on recommendation systems and historical product data (behavioral characteristics) for product recommendations. However, since new products lack behavioral characteristics such as ordering, cancellation, repurchase, and continuous ordering, using existing recommendation methods results in low recommendation accuracy.

[0004] Therefore, how to accurately recommend new products to potential users when order data for new products is lacking is a technical problem that urgently needs to be solved. Summary of the Invention

[0005] In view of the above, it is necessary to provide a product recommendation method, the purpose of which is to achieve effective recommendation of new products when order data is lacking.

[0006] The product recommendation method provided by this invention includes:

[0007] A first tag for the first product is constructed based on the first attribute information of the first product to be recommended. At least one second product with historical order data is obtained based on the first attribute information. A second tag for the second product is constructed based on the second attribute information of the second product.

[0008] Calculate a first similarity value between the first vector representation of the first tag and the second vector representation of the second tag. If the first similarity value is greater than a first preset threshold, determine the list of users to be recommended for the first product based on the order data of the second product.

[0009] Obtain behavioral data of each user in the list of users to be recommended within a preset time period, determine whether the behavioral data contains access data of a third product related to the first attribute information, and if so, construct a third tag for the third product based on the access data.

[0010] Calculate a second similarity value between the first vector representation of the first tag and the third vector representation of the third tag. Based on the second similarity value, select users who meet the preset conditions from the list of users to be recommended as target recommended users, and recommend the first product to the target recommended users.

[0011] Optionally, constructing the first tag of the first product based on the first attribute information of the first product to be recommended includes:

[0012] Select the attribute information that is relevant to user decision-making from all the attribute information of the first product as the first attribute information of the first product;

[0013] The first label of the first product is obtained based on the first attribute information.

[0014] Optionally, before obtaining at least one second product with historical order data based on the first attribute information, the method further includes:

[0015] Collect a preset number of second products with historical order data, and write the second attribute information of the second products into the attribute information record table;

[0016] The attribute information record table is associated with the second product and stored in a preset sales database.

[0017] Optionally, obtaining at least one second product with historical order data based on the first attribute information, and constructing a second tag for the second product based on the second attribute information of the second product, includes:

[0018] The first attribute information is used as a query statement for the attribute information record table of the second product, and at least one second product with historical order data is obtained from the sales database.

[0019] The second label of the second product is obtained based on the second attribute information of the second product.

[0020] Optionally, determining the list of users to be recommended for the first product based on the order data of the second product includes:

[0021] Based on the order data of the second product, obtain the user information of users who have purchased the second product, and summarize the user information to obtain the list of users to be recommended for the first product.

[0022] Optionally, the access data includes:

[0023] Each user to be recommended has at least one of the following: historical browsing information, search history, product collection history, and purchase history of the third product within a preset time period.

[0024] Optionally, the step of selecting users who meet preset conditions from the list of users to be recommended as target recommended users based on the second similarity value includes:

[0025] Users whose second similarity value is greater than the second preset threshold are selected from the list of users to be recommended as the target recommended users.

[0026] To address the above problems, the present invention also provides a product recommendation device, the device comprising:

[0027] The construction module is used to construct a first tag for the first product based on the first attribute information of the first product to be recommended, obtain at least one second product with historical order data based on the first attribute information, and construct a second tag for the second product based on the second attribute information of the second product.

[0028] The calculation module is used to calculate a first similarity value between the first vector representation of the first tag and the second vector representation of the second tag. If the first similarity value is greater than a first preset threshold, the module determines the list of users to be recommended for the first product based on the order data of the second product.

[0029] The judgment module is used to obtain the behavior data of each user to be recommended in the list of users to be recommended within a preset time period, and to determine whether the behavior data contains access data of a third product related to the first attribute information. If so, the third tag of the third product is constructed based on the access data.

[0030] The recommendation module is used to calculate a second similarity value between the first vector representation of the first tag and the third vector representation of the third tag, and to select users who meet preset conditions from the list of users to be recommended as target recommended users based on the second similarity value, and to recommend the first product to the target recommended users.

[0031] To address the above problems, the present invention also provides an electronic device, the electronic device comprising:

[0032] At least one processor; and,

[0033] A memory communicatively connected to the at least one processor; wherein,

[0034] The memory stores a product recommendation program that can be executed by the at least one processor, which enables the at least one processor to perform the product recommendation method described above.

[0035] To address the aforementioned problems, the present invention also provides a computer-readable storage medium storing a product recommendation program, which can be executed by one or more processors to implement the aforementioned product recommendation method.

[0036] Compared with existing technologies, this invention, when the first product is launched and released, can more accurately understand and characterize the characteristics of the first product by constructing a first tag for the first product, thereby improving the accuracy of the recommendation algorithm.

[0037] By calculating the vector representations of the attribute information of the first and second products (new and old products), we can find second products with similar characteristics to the first product and a list of potential recommended users, and thus use the sales data of the second product to guide the recommendation of the first product.

[0038] By analyzing the behavioral data of each user in the recommended user list within a preset time period, we can gain a deeper understanding of their interests and preferences. By determining whether the behavioral data contains access data for a third product related to the first attribute information, we can ascertain the user's interest in the first product and reduce the likelihood of recommending new products that the user is not interested in.

[0039] By calculating the second similarity value of the vector representations corresponding to the attribute information of the first product and the third product, users who meet the preset conditions are selected from the list of users to be recommended based on the second similarity value as target recommended users. The first product is then recommended to the target recommended users. By reducing invalid recommendations, the accuracy of recommendations is improved, thereby ensuring that new products can be accurately recommended to users even when order data for new products is lacking.

[0040] This invention selects a second product with similar characteristics to the first product, uses the sales data of the second product to guide the recommendation of the first product, thus achieving the first layer of protection for the accurate recommendation of the first product. Then, it determines the recommended user list from the second product, and selects the target recommended users who are better suited to recommend the first product from the list of users to be recommended. By reducing invalid recommendations, it achieves the second layer of protection for the accurate recommendation of the first product. Attached Figure Description

[0041] Figure 1 This is a schematic flowchart of a product recommendation method provided in an embodiment of the present invention;

[0042] Figure 2 This is a schematic diagram of a product recommendation device provided in an embodiment of the present invention;

[0043] Figure 3 This is a schematic diagram of the structure of an electronic device for implementing a product recommendation method according to an embodiment of the present invention;

[0044] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0045] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without inventive effort are within the scope of protection of this invention.

[0046] It should be noted that the descriptions involving "first," "second," etc., in this invention are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined with "first" or "second" may explicitly or implicitly include at least one of that feature. Furthermore, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. If the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed by this invention.

[0047] Reference Figure 1 The diagram shown is a schematic flowchart of a product recommendation method provided in an embodiment of the present invention. This method is executed by an electronic device.

[0048] In this embodiment, the product recommendation method includes:

[0049] S1. Construct a first tag for the first product based on the first attribute information of the first product to be recommended, obtain at least one second product with historical order data based on the first attribute information, and construct a second tag for the second product based on the second attribute information of the second product.

[0050] In this embodiment, after a new product (the first product to be recommended) is developed on the financial platform, the first product is input into the product recommendation system, and the type of the first product is automatically identified. The first product includes, but is not limited to, wealth management products and insurance products. The type of the first product includes, but is not limited to, text type, image type and image-text hybrid type.

[0051] Extract attribute information of the first product based on its type. For example, when the first product is a text-type product, natural language processing technology is used to extract keywords contained in the text as multiple attribute information of the first product.

[0052] When the first product is an image-based product, key visual elements in the image (e.g., icons, colors, patterns, numbers, and layouts) are identified using image processing technology as multiple attribute information of the first product.

[0053] When the first product is a mixed image and text product, key visual elements in the image of the first product are identified through image processing technology, and keywords in the text of the first product are extracted through natural language processing technology, which serve as multiple attribute information of the first product.

[0054] Extracting relevant attribute information based on the type of the first product can help us better understand its characteristics, thereby improving the accuracy of recommending the first product to potential customers.

[0055] From all the attribute information of the first product, select the attribute information that is relevant to the user's decision-making as the first attribute information of the first product. Tag the first attribute information to obtain the first tag of the first product. For example, the first product is financial product A, and its first attribute information includes expected return rate of 5%, low risk level, and investment period of 30 days. Tagging the first attribute information to obtain the first tag, which corresponds to "high expected return rate", "low risk", "short term", etc.

[0056] Attributes relevant to user decision-making refer to product characteristics or attributes that can significantly influence users' purchasing decisions. Taking financial products or insurance products as examples, the attributes that influence user decisions include rate of return, risk level, investment period, minimum investment amount, liquidity, fees, and handling charges.

[0057] Tag processing refers to the process of transforming extracted attribute information into structured tags. The main purpose is to represent key product attributes in an easily understandable and processable way, facilitating subsequent similarity calculations and recommendation operations.

[0058] Based on the primary attribute information of the primary product, at least one secondary product with historical order data is retrieved from the sales database. Historical order data refers to a collection of customer purchase behavior data recorded by the company over a period of time. This data collection includes detailed information such as the customer's personal information, order time, purchased goods or services, quantity, price, payment method, and delivery address.

[0059] A sales database is a database used by a company to store data related to sales activities. This data typically includes information about products, customers, orders, transaction records, and so on.

[0060] The second attribute information of the second product is tagged to obtain the second tag of the second product. The first product and the second product have the same or similar attribute information. For example, the first product is financial product A, whose first attribute information includes expected return of 5%, low risk level, and investment period of 30 days. Tagging the first attribute information yields the first tag, which corresponds to "high expected return", "low risk", "short term", etc. Then the second product is financial product B, whose second attribute information includes expected return of 5.6%, low risk level, and investment period of 90 days. Tagging the second attribute information yields the second tag, which corresponds to "high expected return", "low risk", "medium term", etc.

[0061] By constructing a primary label, the characteristics of the primary product can be more accurately understood and represented, thereby improving the accuracy of the recommendation algorithm. When the primary product and the secondary product are similar in some key attribute information, historical order data of the secondary product can be used to provide an appropriate recommendation strategy for the primary product, thereby increasing the recommendation accuracy of the primary product.

[0062] In one embodiment, constructing the first tag of the first product based on the first attribute information of the first product to be recommended includes:

[0063] Select the attribute information that is relevant to user decision-making from all the attribute information of the first product as the first attribute information of the first product;

[0064] The first label of the first product is obtained based on the first attribute information.

[0065] Select the attribute information (such as expected rate of return and risk level) that is relevant to the user's decision-making from the attribute information of the first product (the attribute information includes product name, product code, price, function, expected rate of return, risk level and investment period, etc.) as the first attribute information of the first product, and mark the first attribute information to obtain the first label.

[0066] In one embodiment, before constructing the first label for the first product based on the first attribute information of the first product to be recommended, the method further includes:

[0067] Define all attribute information of the first product according to the type of the first product.

[0068] In one embodiment, defining all attribute information of the first product according to its type includes:

[0069] When the first product is a text-based product, keywords in the text are extracted using natural language processing technology as attribute information for the first product.

[0070] In one embodiment, defining all attribute information of the first product according to its type includes:

[0071] When the first product is an image-based product, key visual elements of the first product are identified using image processing technology as the attribute information of the first product.

[0072] In one embodiment, defining all attribute information of the first product according to its type includes:

[0073] When the first product is a mixed image and text product, key visual elements in the image of the first product are identified using image processing technology as attribute information of the first product, and keywords in the text of the first product are extracted using natural language processing technology as attribute information of the first product.

[0074] In one embodiment, before obtaining at least one second product with historical order data based on the first attribute information, the method further includes:

[0075] Collect a preset number of second products with historical order data, and write the second attribute information of the second products into the attribute information record table;

[0076] The attribute information record table is associated with the second product and stored in a preset sales database.

[0077] Collect a preset number (e.g., 5000) of second products with historical order data, record the second attribute information of each second product in an attribute information record table, associate the attribute information record table with the corresponding second product and store it in the sales database.

[0078] The attribute information record table allows for easy viewing and comparison of the attribute information of the first product, thereby providing a better understanding of the characteristics and market performance of the first product to be recommended.

[0079] In one embodiment, obtaining at least one second product with historical order data based on the first attribute information, and constructing a second tag for the second product based on the second attribute information of the second product, includes:

[0080] The first attribute information is used as a query statement for the attribute information record table of the second product, and at least one second product with historical order data is obtained from the sales database.

[0081] The second label of the second product is obtained based on the second attribute information of the second product.

[0082] The first attribute information of the first product is written into an SQL query statement and used as a query statement for the attribute information record table of the second product. The second product with the same or similar first attribute information as the first product is found in the sales database.

[0083] The secondary attribute information of the second product is labeled to obtain the secondary label of the second product. For example, the second product is wealth management product B, whose secondary attribute information includes expected return of 5.6%, low risk level, and investment period of 90 days. The secondary attribute information is labeled to obtain the secondary label, which corresponds to "high expected return", "low risk", "medium term", etc.

[0084] S2. Calculate the first similarity value between the first vector representation of the first tag and the second vector representation of the second tag. If the first similarity value is greater than a first preset threshold, determine the list of users to be recommended for the first product based on the order data of the second product.

[0085] In this embodiment, the first label of the first product is converted into a first vector representation, and the second label of the second product is converted into a second vector representation. Converting the first and second labels into vector form facilitates the computer program's recognition and processing of these labels, and can quickly improve the efficiency of similarity calculation. For example, the first label [high expected return, low risk, short term] is converted into vector form [1,0,0] using one-hot encoding.

[0086] Calculate the similarity between the first vector representation and the second vector representation to obtain the first similarity value between the first product and the second product.

[0087] If the first similarity value is greater than the first preset threshold (for example, if the first similarity value is 0.6 and the first preset threshold is 0.5, then 0.6 is greater than 0.5), it indicates that the first product and the second product are very similar in attribute information. Therefore, it can be inferred that the first product will be favored by the same user group as the second product. By setting the first preset threshold, the possibility of recommending products that do not match the user's preferences can be reduced, thus establishing the first line of defense against invalid recommendations of the first product.

[0088] Obtain the order data of each user who has already purchased the second product. Since these users have already shown interest in and purchased the second product, which has the same or similar attributes as the first product, there is a high probability that these users will also show interest in or purchase the first product. This will result in a list of users to be recommended for the first product.

[0089] In other embodiments, users who have purchased the second product are screened based on the second attribute information of the first product (i.e., the characteristics of the target user group, including age, gender, occupation, income level, etc.). After screening, it can be determined which users who have purchased the second product meet the audience characteristics of the first product, and these users are listed as potential recommendation targets for the first product, thus obtaining the list of users to be recommended for the first product.

[0090] By calculating the first similarity value between the first product and the second product, we can more accurately identify users who are interested in the first product and recommend it to them. This reduces the need to send recommendation information to users who are not interested and also reduces the marketing costs of new products.

[0091] In one embodiment, calculating the first similarity value between the first vector representation of the first label and the second vector representation of the second label includes:

[0092] Convert the first label of the first product into a first vector representation, and convert the second label of the second product into a second vector representation;

[0093] Calculate the similarity between the first vector representation and the second vector representation to obtain the first similarity value.

[0094] By vectorizing the first label of the first product and the second label of the second product, the first label of the second product is converted into vector form, enabling computer programs to recognize and process the vector form and calculate the similarity.

[0095] Use one or more similarity calculation methods (such as cosine similarity, Euclidean distance, etc.) to calculate the similarity between the first vector representation and the second vector representation.

[0096] Calculating the first similarity value can determine the degree of similarity between the first product and the second product in terms of attribute information, thereby determining whether the first product can be recommended to the users who purchased the second product.

[0097] In one embodiment, determining the list of users to be recommended for the first product based on the order data of the second product includes:

[0098] Based on the order data of the second product, obtain the user information of users who have purchased the second product, and summarize the user information to obtain the list of users to be recommended for the first product.

[0099] The system retrieves order data for the second product from the sales database, extracts information on all users who have purchased the second product, and summarizes all user information into a pre-defined table to obtain a list of users to be recommended for the first product.

[0100] In other embodiments, users who have purchased the second product are filtered based on the second attribute information of the first product (i.e., characteristics of the target user group, including age, gender, occupation, income level, etc.). By obtaining a list of users to be recommended through filtering, it is possible to determine which users who have purchased the second product are more likely to be interested in the first product, thereby improving the effectiveness and conversion rate of the recommendation.

[0101] S3. Obtain the behavior data of each user in the list of users to be recommended within a preset time period, and determine whether the behavior data contains access data of a third product related to the first attribute information. If so, construct a third tag for the third product based on the access data.

[0102] In this embodiment, behavioral data of each user to be recommended is acquired within a preset time period (e.g., the past 3 months or 3 weeks). This includes, but is not limited to, historical browsing information, search records, product collection records, and purchase records. By analyzing the behavioral data of each user to be recommended, the user's personal preferences and behavioral patterns can be effectively identified, thereby improving the accuracy of the recommendations.

[0103] The process involves determining whether the behavioral data of the user to be recommended contains access data for a third product related to the first attribute information. If the user has accessed a third product, the analysis focuses on their browsing, searching, saving, and purchasing data to identify the attributes most frequently used by the user as the third attribute information for the third product. For example, if the user frequently uses keywords such as "high expected rate of return" and "low risk" when searching for a third product, then "high expected rate of return" and "low risk" are considered third attribute information for the third product. This third attribute information is then tagged to obtain the third tags for the third product.

[0104] Determining whether the behavioral data of the user to be recommended contains access data of a third product related to the first attribute information can reduce the possibility of recommending products that the user is not interested in, thus establishing a second line of defense against invalid recommendations of the first product.

[0105] In one embodiment, the access data includes:

[0106] Each user to be recommended has at least one of the following: historical browsing information, search history, product collection history, and purchase history of the third product within a preset time period.

[0107] Historical browsing information refers to a record of the web pages or third-party products that a user to be recommended has viewed over a past period. By analyzing the browsing history of a user to be recommended, we can understand the types of products or topics that the user is interested in.

[0108] Search history refers to the keywords or phrases that the user to be recommended has searched for on a website or application over a period of time. Search history can help understand the specific needs or interests of the user to be recommended.

[0109] "Favorited products" refers to products that the user to be recommended has added to their favorites or wishlist. Favorited products typically reflect the product types or brands that the user has been following in the long term.

[0110] Purchase history refers to the records of third-party products purchased by the user to be recommended over a period of time. Purchase history is one of the most direct sources of data reflecting the preferences of the user to be recommended, and can be used to predict products that the user may be interested in in the future.

[0111] Behavioral data comprises a series of online behavioral records of the user to be recommended within a preset time period. These records help the recommendation system better understand the user's interests, preferences, and purchase intentions. By analyzing this data, the product recommendation system can more accurately recommend products that match the user's interests, thereby improving recommendation accuracy and user satisfaction.

[0112] In one embodiment, constructing the third tag for the third product based on the access data includes:

[0113] The attribute information of the third product that is related to the user's decision is obtained as the third attribute information of the third product.

[0114] The third attribute information is processed to obtain the third tag of the third product.

[0115] Analyzing the behavioral data of users to be recommended within a preset time period, such as browsing history, search records, product favorites, and purchase history, helps identify information from all the attributes of the third-party product that significantly influences user decisions. For example, if users frequently use keywords like "high expected return" and "low risk" in their search history, these keywords are likely to have a significant impact on their decisions. These keywords are then used as the third attribute information of the third-party product, and the selected third attribute information is labeled to obtain the third tag of the third-party product. This allows for a more accurate capture of the true needs and interests of users to be recommended, thereby improving the accuracy of recommendations.

[0116] S4. Calculate the second similarity value between the first vector representation of the first tag and the third vector representation of the third tag, and select users who meet the preset conditions from the list of users to be recommended as target recommended users based on the second similarity value, and recommend the first product to the target recommended users.

[0117] In this embodiment, the first label of the first product is converted into a first vector representation, and the second label of the third product is converted into a third vector representation. Converting the first and second labels into vector form facilitates the computer program's recognition and processing of these labels, and enables efficient and rapid calculation of similarity. The similarity between the first and third vector representations is calculated to obtain a second similarity value between the first and third products.

[0118] By calculating the second similarity value between the first and third products, we can more accurately identify target users who are interested in the first product and reduce the need to send recommendation information to users who are not interested.

[0119] If the second similarity value is greater than the second preset threshold (for example, if the second similarity value is 0.6 and the second preset threshold is 0.5, then 0.6 is greater than 0.5), it indicates that the first product and the third product are very similar in attribute information. Therefore, it can be concluded that the first product will be favored by the same user group as the third product. By setting a second preset threshold, the possibility of recommending products that do not match the user's preferences can be reduced, thus establishing a third line of protection against invalid recommendations of the first product.

[0120] The implementation of the present invention is illustrated by the following examples:

[0121] Suppose a financial platform offers a new wealth management product A (Product 1). The platform wants to recommend it to suitable users and has found a wealth management product B (Product 2) with a certain sales record from its sales database. Both products have the characteristics of high expected returns and low risk. Obtaining the user group that purchased wealth management product B will give the list of users to be recommended to Product 1.

[0122] Analyze the behavioral data of each user in the list of users to be recommended to see if they have visited other financial products (third products) with similar attributes to financial product A. For example, if user K previously visited another low-risk, high-expected-return financial product C (third product), then user K can be considered to be interested in financial product A. Based on the key information such as "low risk" and "high expected return" used by user K when visiting financial product C, construct corresponding tags for financial product C, such as "high expected return" and "low risk".

[0123] Calculate the second similarity value between the first tag of financial product A and the third tag of financial product C. Assuming the second similarity value is 0.6 and the preset threshold is 0.5, the recommendation criteria for financial product A are met. Detailed information about financial product A will then be pushed to users who have accessed financial product C via email, SMS, or in-app notifications.

[0124] In one embodiment, calculating a second similarity value between the first vector representation of the first label and the third vector representation of the third label includes:

[0125] Convert the third label of the third product into a third vector representation;

[0126] Calculate the similarity between the first vector representation and the third vector representation to obtain the second similarity value.

[0127] By vectorizing the first label of the first product and the third label of the third product into vector form, the computer program can recognize and process the vector form and calculate the similarity.

[0128] Use one or more similarity calculation methods (such as cosine similarity, Euclidean distance, etc.) to calculate the similarity between the first vector representation and the third vector representation.

[0129] Calculating the second similarity value can determine the degree of similarity between the first product and the third product in terms of attribute information, thereby filtering out the final target recommended users from the list of users to be recommended, and recommending the first product to the target recommended users.

[0130] In one embodiment, the step of filtering users who meet preset conditions from the list of users to be recommended as target recommended users based on the second similarity value includes:

[0131] Users whose second similarity value is greater than the second preset threshold are selected from the list of users to be recommended as the target recommended users.

[0132] In one embodiment, recommending the first product to the target recommended user includes:

[0133] Obtain the recommendation method of the target recommended user, and recommend the first product to the target recommended user through the recommendation method.

[0134] The method for obtaining recommendations from target users can be any way that sends or displays the primary product to users. Examples include email, SMS, in-app notifications, or personalized recommendation interfaces.

[0135] The first product is recommended to the target users, along with a detailed description of the product, its features, and advantages, so that users can understand the product's details and decide whether to purchase. By filtering target users from the list of users to be recommended, the first product is ensured to be recommended to those most likely to be interested, thereby improving the accuracy and effectiveness of the recommendations and contributing to higher sales conversion rates.

[0136] In steps S1-S4 above, when the first product is launched, by constructing the first tag of the first product, the characteristics of the first product can be understood and represented more accurately, thereby improving the accuracy of the recommendation algorithm.

[0137] By calculating the vector representations of the attribute information of the first and second products (new and old products), we can find second products with similar characteristics to the first product and a list of potential recommended users, and thus use the sales data of the second product to guide the recommendation of the first product.

[0138] By analyzing the behavioral data of each user in the recommended user list within a preset time period, we can gain a deeper understanding of their interests and preferences. By determining whether the behavioral data contains access data for a third product related to the first attribute information, we can ascertain the user's interest in the first product and reduce the likelihood of recommending new products that the user is not interested in.

[0139] By calculating the second similarity value of the vector representations corresponding to the attribute information of the first product and the third product, users who meet the preset conditions are selected from the list of users to be recommended based on the second similarity value as target recommended users. The first product is then recommended to the target recommended users. By reducing invalid recommendations, the accuracy of recommendations is improved, thereby ensuring that new products can be accurately recommended to users even when order data for new products is lacking.

[0140] This invention selects a second product with similar characteristics to the first product, uses the sales data of the second product to guide the recommendation of the first product, thus achieving the first layer of protection for the accurate recommendation of the first product. Then, it determines the recommended user list from the second product, and selects the target recommended users who are better suited to recommend the first product from the list of users to be recommended. By reducing invalid recommendations, it achieves the second layer of protection for the accurate recommendation of the first product.

[0141] like Figure 2 The diagram shown is a schematic representation of a product recommendation device according to an embodiment of the present invention.

[0142] The product recommendation device 100 of this invention can be installed in an electronic device. Depending on the functions implemented, the product recommendation device 100 may include a construction module 110, a calculation module 120, a judgment module 130, and a recommendation module 140. The module described in this invention can also be called a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, and which are stored in the memory of the electronic device.

[0143] In this embodiment, the functions of each module / unit are as follows:

[0144] The construction module 110 is used to construct a first tag for the first product based on the first attribute information of the first product to be recommended, obtain at least one second product with historical order data based on the first attribute information, and construct a second tag for the second product based on the second attribute information of the second product.

[0145] The calculation module 120 is used to calculate a first similarity value between the first vector representation of the first tag and the second vector representation of the second tag. If the first similarity value is greater than a first preset threshold, the list of users to be recommended for the first product is determined based on the order data of the second product.

[0146] The judgment module 130 is used to obtain the behavior data of each user to be recommended in the list of users to be recommended within a preset time period, and to determine whether the behavior data contains access data of a third product related to the first attribute information. If so, the third tag of the third product is constructed based on the access data.

[0147] The recommendation module 140 is used to calculate a second similarity value between the first vector representation of the first tag and the third vector representation of the third tag, and to select users who meet preset conditions from the list of users to be recommended as target recommended users based on the second similarity value, and to recommend the first product to the target recommended users.

[0148] In one embodiment, constructing the first tag of the first product based on the first attribute information of the first product to be recommended includes:

[0149] Select the attribute information that is relevant to user decision-making from all the attribute information of the first product as the first attribute information of the first product;

[0150] The first label of the first product is obtained based on the first attribute information.

[0151] In one embodiment, before obtaining at least one second product with historical order data based on the first attribute information, the method further includes:

[0152] Collect a preset number of second products with historical order data, and write the second attribute information of the second products into the attribute information record table;

[0153] The attribute information record table is associated with the second product and stored in a preset sales database.

[0154] In one embodiment, obtaining at least one second product with historical order data based on the first attribute information, and constructing a second tag for the second product based on the second attribute information of the second product, includes:

[0155] The first attribute information is used as a query statement for the attribute information record table of the second product, and at least one second product with historical order data is obtained from the sales database.

[0156] The second label of the second product is obtained based on the second attribute information of the second product.

[0157] In one embodiment, determining the list of users to be recommended for the first product based on the order data of the second product includes:

[0158] Based on the order data of the second product, obtain the user information of users who have purchased the second product, and summarize the user information to obtain the list of users to be recommended for the first product.

[0159] In one embodiment, the access data includes:

[0160] Each user to be recommended has at least one of the following: historical browsing information, search history, product collection history, and purchase history of the third product within a preset time period.

[0161] In one embodiment, the step of filtering users who meet preset conditions from the list of users to be recommended as target recommended users based on the second similarity value includes:

[0162] Users whose second similarity value is greater than the second preset threshold are selected from the list of users to be recommended as the target recommended users.

[0163] like Figure 3 The diagram shown is a structural schematic of an electronic device for implementing a product recommendation method according to an embodiment of the present invention.

[0164] In this embodiment, the electronic device 1 includes, but is not limited to, a memory 11, a processor 12, and a network interface 13 that can be interconnected via a system bus. The memory 11 stores a product recommendation program 10, which can be executed by the processor 12. Figure 3 Only the electronic device 1 with components 11-13 and product recommendation program 10 is shown. Those skilled in the art will understand that... Figure 3 The structure shown does not constitute a limitation on the electronic device 1, and may include fewer or more components than shown, or combine certain components, or have different component arrangements.

[0165] The memory 11 includes RAM and at least one type of readable storage medium. The RAM provides a cache for the operation of the electronic device 1; the readable storage medium can be a non-volatile storage medium such as flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the readable storage medium can be an internal storage unit of the electronic device 1; in other embodiments, the non-volatile storage medium can also be an external storage device of the electronic device 1, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the electronic device 1. In this embodiment, the readable storage medium of the memory 11 is typically used to store the operating system and various application software installed on the electronic device 1, such as storing the code of the product recommendation program 10 in one embodiment of the present invention. Furthermore, the memory 11 can also be used to temporarily store various types of data that have been output or will be output.

[0166] In some embodiments, processor 12 may be a central processing unit (CPU), controller, microcontroller, microprocessor, or other data processing chip. Processor 12 is typically used to control the overall operation of the electronic device 1, such as performing control and processing related to data interaction or communication with other devices. In this embodiment, processor 12 is used to run program code stored in memory 11 or process data, such as running product recommendation program 10.

[0167] The network interface 13 may include a wireless network interface or a wired network interface, which is used to establish a communication connection between the electronic device 1 and the terminal (not shown in the figure).

[0168] Optionally, the electronic device 1 may further include a user interface, which may include a display, an input unit such as a keyboard, and optionally, a standard wired interface or a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen, etc. The display may also be appropriately referred to as a screen or display unit, used to display information processed in the electronic device 1 and to display a visual user interface.

[0169] It should be understood that the embodiments described are for illustrative purposes only and are not limited to this structure in the scope of the patent application.

[0170] The product recommendation program 10 stored in the memory 11 of the electronic device 1 is a combination of multiple instructions, which, when run in the processor 12, can achieve the following:

[0171] A first tag for the first product is constructed based on the first attribute information of the first product to be recommended. At least one second product with historical order data is obtained based on the first attribute information. A second tag for the second product is constructed based on the second attribute information of the second product.

[0172] Calculate a first similarity value between the first vector representation of the first tag and the second vector representation of the second tag. If the first similarity value is greater than a first preset threshold, determine the list of users to be recommended for the first product based on the order data of the second product.

[0173] Obtain behavioral data of each user in the list of users to be recommended within a preset time period, determine whether the behavioral data contains access data of a third product related to the first attribute information, and if so, construct a third tag for the third product based on the access data.

[0174] Calculate a second similarity value between the first vector representation of the first tag and the third vector representation of the third tag. Based on the second similarity value, select users who meet the preset conditions from the list of users to be recommended as target recommended users, and recommend the first product to the target recommended users.

[0175] Specifically, the processor 12's implementation method for the aforementioned product recommendation program 10 can be found in [reference needed]. Figure 1 The descriptions of the relevant steps in the corresponding embodiments are not repeated here.

[0176] Furthermore, if the modules / units integrated in the electronic device 1 are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. The computer-readable medium can be non-volatile or non-combustible. The computer-readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, or a read-only memory (ROM).

[0177] The computer-readable storage medium stores a product recommendation program 10, which can be executed by one or more processors. The specific implementation of the computer-readable storage medium of the present invention is basically the same as the embodiments of the product recommendation method described above, and will not be repeated here.

[0178] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0179] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0180] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0181] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0182] Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be embraced within the invention. No appended diagram markings in the claims should be construed as limiting the scope of the claims.

[0183] Furthermore, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices recited in a system claim may also be implemented by a single unit or device through software or hardware. The term "second class" is used to indicate names and does not indicate any specific order.

[0184] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A product recommendation method, characterized in that, The method includes: A first tag for the first product is constructed based on the first attribute information of the first product to be recommended. At least one second product with historical order data is obtained based on the first attribute information. A second tag for the second product is constructed based on the second attribute information of the second product. Calculate a first similarity value between the first vector representation of the first tag and the second vector representation of the second tag. If the first similarity value is greater than a first preset threshold, determine the list of users to be recommended for the first product based on the order data of the second product. Obtain behavioral data of each user in the list of users to be recommended within a preset time period, determine whether the behavioral data contains access data of a third product related to the first attribute information, and if so, construct a third tag for the third product based on the access data. Calculate a second similarity value between the first vector representation of the first tag and the third vector representation of the third tag, and select users who meet the preset conditions from the list of users to be recommended as target recommended users based on the second similarity value, and recommend the first product to the target recommended users; Before obtaining at least one second product with historical order data based on the first attribute information, the method further includes: collecting a preset number of second products with historical order data; writing the second attribute information of the second products into an attribute information record table; associating the attribute information record table with the second products and storing it in a preset sales database; The step of obtaining at least one second product with historical order data based on the first attribute information and constructing a second tag for the second product based on the second attribute information of the second product includes: using the first attribute information as a query statement for the attribute information record table of the second product to query at least one second product with historical order data from the sales database; and obtaining a second tag for the second product based on the second attribute information of the second product.

2. The product recommendation method as described in claim 1, characterized in that, The step of constructing the first tag of the first product based on the first attribute information of the first product to be recommended includes: Select the attribute information that is relevant to user decision-making from all the attribute information of the first product as the first attribute information of the first product; The first label of the first product is obtained based on the first attribute information.

3. The product recommendation method as described in claim 1, characterized in that, The step of determining the list of users to be recommended for the first product based on the order data of the second product includes: Based on the order data of the second product, obtain the user information of users who have purchased the second product, and summarize the user information to obtain the list of users to be recommended for the first product.

4. The product recommendation method as described in claim 1, characterized in that, The access data includes: Each user to be recommended has at least one of the following: historical browsing information, search history, product collection history, and purchase history of the third product within a preset time period.

5. The product recommendation method as described in claim 1, characterized in that, The step of selecting users who meet preset conditions from the list of users to be recommended as target recommended users based on the second similarity value includes: Users whose second similarity value is greater than the second preset threshold are selected from the list of users to be recommended as the target recommended users.

6. A product recommendation device, characterized in that, The device includes: The construction module is used to construct a first tag for the first product based on the first attribute information of the first product to be recommended, obtain at least one second product with historical order data based on the first attribute information, and construct a second tag for the second product based on the second attribute information of the second product. The calculation module is used to calculate a first similarity value between the first vector representation of the first tag and the second vector representation of the second tag. If the first similarity value is greater than a first preset threshold, the module determines the list of users to be recommended for the first product based on the order data of the second product. The judgment module is used to obtain the behavior data of each user to be recommended in the list of users to be recommended within a preset time period, and to determine whether the behavior data contains access data of a third product related to the first attribute information. If so, the third tag of the third product is constructed based on the access data. The recommendation module is used to calculate a second similarity value between the first vector representation of the first tag and the third vector representation of the third tag, and to select users who meet preset conditions from the list of users to be recommended as target recommended users based on the second similarity value, and to recommend the first product to the target recommended users. Before obtaining at least one second product with historical order data based on the first attribute information, the device is further configured to: collect a preset number of second products with historical order data; write the second attribute information of the second products into an attribute information record table; associate the attribute information record table with the second products and store it in a preset sales database; The step of obtaining at least one second product with historical order data based on the first attribute information and constructing a second tag for the second product based on the second attribute information of the second product includes: using the first attribute information as a query statement for the attribute information record table of the second product to query at least one second product with historical order data from the sales database; and obtaining a second tag for the second product based on the second attribute information of the second product.

7. An electronic device, characterized in that, The electronic device includes: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores a product recommendation program that can be executed by the at least one processor, the product recommendation program being executed by the at least one processor to enable the at least one processor to perform the product recommendation method as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a product recommendation program, which can be executed by one or more processors to implement the product recommendation method as described in any one of claims 1 to 5.

Images