A combination product recommendation method and device, electronic equipment and storage medium
By acquiring user characteristics and consumption data, and utilizing pre-defined customer groups and machine learning models, a combined product recommendation scheme is determined, which solves the problems of low recommendation rationality and accuracy in existing technologies and achieves more efficient combined product recommendation.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-11-17
- Publication Date
- 2026-04-07
AI Technical Summary
Existing product recommendation schemes consider too few dimensions, resulting in low rationality and accuracy of recommendations.
By acquiring the current user's characteristic data and consumption data, and using a pre-defined customer group and machine learning model, the initial recommended product combination for the user is determined, and the final recommended product combination is determined based on the purchase probability.
This improved the rationality and accuracy of product combination recommendations, and enhanced their effectiveness.
Smart Images
Figure CN115757956B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of intelligent decision-making technology, and in particular to a method, apparatus, electronic device, and computer-readable storage medium for recommending combined products. Background Technology
[0002] In today's era of advanced internet and big data, many internet companies offer multiple products. For example, in the auto insurance business, when choosing commercial insurance, users may select one or more coverage options from multiple alternatives. However, most users have limited information; therefore, accurately recommending bundled products is highly valuable for increasing sales and operating profits. However, existing bundled product recommendation schemes consider too few dimensions, resulting in lower rationality and accuracy in the recommendations. Summary of the Invention
[0003] The purpose of this invention is to provide a method, apparatus, electronic device, and computer-readable storage medium for recommending combined products, so as to solve the technical problem of low accuracy in the recommendation of combined products in the prior art.
[0004] The technical solution of the present invention is as follows: a method for recommending combined products is provided, comprising:
[0005] Obtain the current user's characteristic data and consumption data;
[0006] Based on the current user's characteristic data, determine the customer group corresponding to the current user in the preset customer group. Based on the customer group corresponding to the current user and the preset customer group, determine the initial recommended product combination corresponding to the current user. The preset customer group is determined based on the characteristic data and consumption data of a preset number of users. The preset customer group includes the customer group to which each user belongs in the preset number of users and the initial recommended product combination corresponding to each customer group.
[0007] The current user's feature data is input into a pre-trained machine learning model to obtain the probability that the current user will purchase different preset products. The machine learning module is trained based on sample feature data and sample consumption data.
[0008] The final recommended product combination is determined based on the initial recommended product combination for the current user and the probability of the current user purchasing different preset products.
[0009] Furthermore, the step of determining the preset customer group includes:
[0010] Obtain feature data and consumption data of a preset number of users;
[0011] The feature data is used to create a profile to obtain user profile data. Based on the user profile data, all customer groups of the preset number of users and the customer group to which each user in the preset number of users belongs are determined.
[0012] Based on the customer group to which each user belongs and the consumption data of each user, an initial recommended product combination is determined for each customer group.
[0013] Further, determining, based on the user profile data, all customer groups of the preset number of users and the customer group to which each user belongs within the preset number of users includes:
[0014] Based on the user profile data, customer group factors and representative attributes of the customer group factors are determined. Based on the customer group factors and representative attributes of the customer group factors, all customer groups of the preset number of users and the customer group to which each user in the preset number of users belongs are determined.
[0015] Furthermore, based on the current user's characteristic data, the customer group corresponding to the current user is determined from the preset customer groups, including:
[0016] Based on the current user's feature data, determine the customer group factor corresponding to the current user and the representative attribute corresponding to the customer group factor. Compare the similarity between the representative attribute corresponding to the customer group factor and the representative attribute corresponding to the customer group factor of all customer groups. Determine the customer group corresponding to the current user based on the similarity comparison result.
[0017] Further, based on the initial recommended product combination corresponding to the current user and the probability of the current user purchasing different preset products, the final recommended product combination is determined, including:
[0018] If the preset product is a product in the initial recommended product combination, then when the probability of the current user purchasing the preset product is greater than or equal to the first preset probability, the preset product is used as the final recommended product.
[0019] If the preset product is not a product in the initial recommended product combination, then when the probability of the current user purchasing the preset product is greater than or equal to the second preset probability, the preset product will be used as the final recommended product.
[0020] The final recommended product portfolio consists of all the aforementioned final recommended products.
[0021] Furthermore, the current user's consumption data includes the products the current user has purchased in the past and the time of purchase. The method also includes determining pre-recommended products based on the products the current user has purchased in the past and the time of purchase.
[0022] Accordingly, the final recommended product combination is determined based on the initial recommended product combination corresponding to the current user and the probability of the current user purchasing different preset products, including: determining the final recommended product combination based on the initial recommended product combination corresponding to the current user, the pre-recommended products, and the probability of the current user purchasing different preset products.
[0023] Further, based on the initial recommended product combination corresponding to the current user, the pre-recommended products, and the probability of the current user purchasing different preset products, the final recommended product combination is determined, including:
[0024] If the preset product is a product in the initial recommended product combination, or is the pre-recommended product, then when the probability of the current user purchasing the preset product is greater than or equal to the first preset probability, the preset product shall be used as the final recommended product.
[0025] If the preset product is not a product in the initial recommended product combination and is not a pre-recommended product, then when the probability of the current user purchasing the preset product is greater than or equal to the second preset probability, the preset product will be used as the final recommended product.
[0026] The final recommended product portfolio consists of all the aforementioned final recommended products.
[0027] Another technical solution of the present invention is as follows: a combined product recommendation device is provided, including a data acquisition module, an initial recommendation module, a probability acquisition module, and a final recommendation module;
[0028] The data acquisition module is used to acquire the current user's feature data and consumption data;
[0029] The initial recommendation module is used to determine the customer group corresponding to the current user in a preset customer group based on the current user's feature data, and to determine the initial recommended product combination corresponding to the current user based on the customer group corresponding to the current user and the preset customer group. The preset customer group is determined based on the feature data and consumption data of a preset number of users. The preset customer group includes the customer group to which each user belongs in the preset number of users and the initial recommended product combination corresponding to each customer group.
[0030] The probability acquisition module is used to input the current user's feature data into a pre-trained machine learning model to obtain the probability that the current user will purchase different preset products. The machine learning module is trained based on sample feature data and sample consumption data.
[0031] The final recommendation module is used to determine the final recommended product combination based on the initial recommended product combination corresponding to the current user and the probability of the current user purchasing different preset products.
[0032] Another technical solution of the present invention is as follows: an electronic device is provided, including a memory and a processor. The memory stores a computer program that can be executed by the processor. When the processor executes the computer program, it implements the combined product recommendation method as described in any of the above technical solutions.
[0033] Another technical solution of the present invention is as follows: a computer-readable storage medium is provided, the computer-readable storage medium storing a computer program, which, when executed by a processor, implements the combined product recommendation method as described in any of the above technical solutions.
[0034] The beneficial effects of this invention are as follows: It acquires the characteristic data and consumption data of the current user; based on the current user's characteristic data, it determines the customer group corresponding to the current user within a preset customer group; based on the customer group corresponding to the current user and the preset customer group, it determines the initial recommended product combination corresponding to the current user, wherein the preset customer group is determined based on the characteristic data and consumption data of a preset number of users, and the preset customer group includes the customer group to which each user belongs and the initial recommended product combination corresponding to each customer group; it inputs the current user's characteristic data into a pre-trained machine learning model to obtain the probability that the current user will purchase different preset products, wherein the machine learning module is trained based on sample characteristic data and sample consumption data; based on the initial recommended product combination corresponding to the current user and the probability that the current user will purchase different preset products, it determines the final recommended product combination; through the above technical solution, the rationality of the combined product recommendation is improved, thereby improving the accuracy of the combined product recommendation. Attached Figure Description
[0035] Figure 1 This is a flowchart illustrating the combined product recommendation method according to an embodiment of the present invention;
[0036] Figure 2 This is a schematic diagram of the combined product recommendation device according to an embodiment of the present invention;
[0037] Figure 3 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention. Detailed Implementation
[0038] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.
[0039] In the description of this application, the terms "first," "second," etc., are used only for distinguishing purposes and should not be construed as indicating or implying relative importance or order. In this specification, the terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.
[0040] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.
[0041] Figure 1 This is a flowchart illustrating the combined product recommendation method according to an embodiment of the present invention. It should be noted that if substantially the same result is obtained, the combined product recommendation method of the present invention does not necessarily follow the same approach. Figure 1 The illustrated process sequence is limited. For example... Figure 1 As shown, this product recommendation method mainly includes the following steps:
[0042] S1, obtain the current user's characteristic data and consumption data;
[0043] The feature data may include user age, user gender, user income level, zodiac sign, and business indicator data. For example, for auto insurance business, business indicator data may include vehicle age, vehicle purchase price, whether it is a price-sensitive type, and whether it is a renewal user. The consumption data includes the products purchased by the user in the past, which can be products purchased in the past period of time.
[0044] S2, Based on the current user's characteristic data, determine the customer group corresponding to the current user in the preset customer group, and determine the initial recommended product combination corresponding to the current user based on the customer group corresponding to the current user and the preset customer group. The preset customer group is determined based on the characteristic data and consumption data of a preset number of users. The preset customer group includes the customer group to which each user belongs in the preset number of users and the initial recommended product combination corresponding to each customer group.
[0045] The preset number of users can be all users, which refers to all customer groups in the current company business, and the initial recommended product combination can be at least one product.
[0046] S3, input the current user's feature data into a pre-trained machine learning model to obtain the probability that the current user will purchase different preset products, wherein the machine learning module is trained based on sample feature data and sample consumption data;
[0047] This involves using sample feature data and sample consumption data, and employing the sklearn data analysis framework, to model and train machine learning models (such as logistic regression or support vector machine models) for users purchasing different preset products. After training, the machine learning model will determine the probability of a user purchasing a preset product based on the user's feature data and consumption data. Some products may have a cyclical nature, such as car insurance, where policies are typically renewed annually. If a user is still within the policy's validity period, it's not very meaningful for the machine learning model to determine whether the user will purchase the product. Therefore, when modeling and training the machine learning model, the product's characteristics and cyclicality will be considered. For example, for products with annual renewals, the probability of future purchases can be predicted using data from the most recent 12 months.
[0048] S4. Based on the initial recommended product combination for the current user and the probability of the current user purchasing different preset products, determine the final recommended product combination. The final recommended product combination may contain at least one product.
[0049] This invention, in its embodiments, acquires the current user's feature data and consumption data; based on the current user's feature data, determines the customer group corresponding to the current user within a preset customer group; based on the current user's corresponding customer group and the preset customer group, determines the initial recommended product combination corresponding to the current user; inputs the current user's feature data into a pre-trained machine learning model to obtain the probability of the current user purchasing different preset products; and based on the initial recommended product combination corresponding to the current user and the probability of the current user purchasing different preset products, determines the final recommended product combination; this improves the rationality of the product combination recommendation, thereby improving the accuracy of the product combination recommendation.
[0050] In an optional implementation, the step of determining the preset customer group includes:
[0051] Obtain feature data and consumption data of a preset number of users;
[0052] The feature data is used to create a profile to obtain user profile data. Based on the user profile data, all customer groups of the preset number of users and the customer group to which each user in the preset number of users belongs are determined.
[0053] Based on the customer group to which each user belongs and the consumption data of each user, an initial recommended product combination is determined for each customer group.
[0054] The user feature data obtained can include anonymized data or detailed data, which can be used to create a 360-degree user profile.
[0055] In one specific embodiment, the process of creating a user profile from the feature data includes: performing business standard profiling, equal-frequency profiling, or equal-width profiling on the feature data to obtain user profile data. Specifically, some data indicators of the feature data can be profiled using business standards. For example, age data can be profiled using business standards, such as labeling 0-3 years old as the preschool stage, 4-13 years old as the children's stage, etc., to achieve profiling based on business standards (business experience). Alternatively, other data indicators of the feature data can be profiled using equal-frequency profiling, such as labeling the data indicator equally as good, neutral, or bad. Similarly, other data indicators of the feature data can be profiled using equal-width profiling, such as labeling values of 1-3 as small, values of 4-6 as medium, and values of 7-9 as large. These different profiling methods can be flexibly used when creating user profiles.
[0056] In an optional implementation, determining all customer groups of the preset number of users and the customer group to which each user belongs, based on the user profile data, includes:
[0057] Based on the user profile data, customer group factors and representative attributes of the customer group factors are determined. Based on the customer group factors and representative attributes of the customer group factors, all customer groups of the preset number of users and the customer group to which each user in the preset number of users belongs are determined.
[0058] In one specific embodiment, based on the user profile data, customer group factors and their representative attributes are determined. Regional differences can be considered when determining these factors and their representative attributes. The more customer group factors there are, the richer the customer base becomes, and the more complex the customer acquisition process becomes. If a customer group includes 7 customer group factors, with 3 representative attributes for the first factor, 5 for the second, 4 for the third, 3 for the fourth, 5 for the fifth, 5 for the sixth, and 4 for the seventh, then the number of customers in this group is 3*5*4*3*5*5*4 = 18000. It should be noted that if age is a customer group factor, then the attribute representing childhood is the representative attribute.
[0059] In another specific embodiment, based on the customer group factor and the representative attributes of the customer group factor, all customer groups of the preset number of users and the customer group to which each user in the preset number belongs are determined. Based on the customer group to which each user in the preset number belongs, a customer group table can be obtained, as shown in Table 1.
[0060] Table 1 Customer Group Table
[0061] User ID Target customer group Purchase products 252155 Customer Group B A, B, C 262155 Customer Group A A, D 262175 Customer Group B C, D 263175 Customer Group B A, B, C … … …
[0062] In Table 1, the last column represents the products purchased by different users in the past. The initial recommended product combination for the current user is determined based on the customer group corresponding to the current user, the customer group to which each user belongs in the preset number, and the initial recommended product combination corresponding to each customer group. The initial recommended product combination for the customer group is determined by counting the products purchased in the past by all users in the customer group, sorting them, and using the products purchased in the past with the previous preset ranking as the initial recommended product combination. For example, as shown in Table 1, the initial recommended product combination for customer group B is A, B, and C.
[0063] In a specific implementation, a configuration information table can be output using big data. This configuration information table includes the initial recommended product combination for each customer group. If there are new products, product configuration can also be performed through this configuration information table.
[0064] In an optional implementation, based on the current user's characteristic data, the customer group corresponding to the current user is determined from a preset customer group, including:
[0065] Based on the current user's feature data, determine the customer group factor corresponding to the current user and the representative attribute corresponding to the customer group factor. Compare the similarity between the representative attribute corresponding to the customer group factor and the representative attribute corresponding to the customer group factor of all customer groups. Determine the customer group corresponding to the current user based on the similarity comparison result.
[0066] In one specific embodiment, the customer group factor corresponding to the current user needs to be the same as the customer group factors corresponding to all the above-mentioned customer groups. The representative attribute corresponding to the customer group factor is compared with the representative attribute of the customer group factors corresponding to all the above-mentioned customer groups to obtain a similarity comparison result. The similarity comparison result can be a similarity score. If a customer group includes n customer group factors, the formula for calculating the similarity score can be: similarity score = weight of the first customer group factor * x1 + weight of the second customer group factor * x2 + ... + weight of the nth customer group factor * xn. The weights of the first customer group factor to the nth customer group factor can be preset. x1...xn represent 1 or 0 respectively. If the representative attributes of the two customer group factors are the same, then 1 is taken; otherwise, 0 is taken.
[0067] In an optional implementation, the final recommended product combination is determined based on the initial recommended product combination corresponding to the current user and the probability of the current user purchasing different preset products, including:
[0068] If the preset product is a product in the initial recommended product combination, then when the probability of the current user purchasing the preset product is greater than or equal to the first preset probability, the preset product is used as the final recommended product.
[0069] If the preset product is not a product in the initial recommended product combination, then when the probability of the current user purchasing the preset product is greater than or equal to the second preset probability, the preset product will be used as the final recommended product.
[0070] The final recommended product portfolio consists of all the aforementioned final recommended products.
[0071] In one specific embodiment, the feature data of the current user is input into a pre-trained machine learning model to obtain the probability of the current user purchasing different preset products. For example, if there are several preset products (e.g., 5), the probability of the current user purchasing each of the several different preset products can be predicted by the pre-trained machine learning model. The probability value is between 0 and 1. The probability table of the current user purchasing preset products is shown in Table 2.
[0072] Table 2: Probability of Current Users Purchasing Preset Products
[0073] Product Name probability Product A 0.71 Product B 0.42 Product C 0.83
[0074] In a specific embodiment, the second preset probability can be greater than the first preset probability. For example, the first preset probability can be 0.2 and the second preset probability can be 0.5. If the preset product is not a product in the initial recommended product combination, then when the probability of the current user purchasing the preset product is greater than or equal to 0.5, it is determined that the user will purchase the preset product, and the preset product is used as the final recommended product. As shown in Table 2, product A and product C can be used as the final recommended products.
[0075] If the preset product is a product in the initial recommended product combination, then when the probability of the current user purchasing the preset product is greater than or equal to 0.2, it is determined that the user will purchase the preset product, and the preset product is used as the final recommended product.
[0076] In an optional implementation, the current user's consumption data includes the products the current user has historically purchased and the time of purchase. The method further includes determining pre-recommended products based on the current user's historically purchased products and the time of purchase. Correspondingly, determining the final recommended product combination based on the initial recommended product combination corresponding to the current user and the probability of the current user purchasing different preset products includes: determining the final recommended product combination based on the initial recommended product combination corresponding to the current user, the pre-recommended products, and the probability of the current user purchasing different preset products.
[0077] In one specific embodiment, the current user's consumption data includes the products the current user has purchased in the past and the time of purchase. Based on the products the current user has purchased in the past and the time of purchase, it is determined whether to recommend the previously purchased products, that is, to determine the pre-recommended products. For example, if the user has purchased product A in the past, but the current time is greater than or the recommendation time of product A, it is assumed that the user is currently using it and product A is not recommended. Otherwise, product A is recommended and product A is the pre-recommended product.
[0078] In an optional implementation, the final recommended product combination is determined based on the initial recommended product combination corresponding to the current user, the pre-recommended products, and the probability of the current user purchasing different preset products, including:
[0079] If the preset product is a product in the initial recommended product combination, or is the pre-recommended product, then when the probability of the current user purchasing the preset product is greater than or equal to the first preset probability, the preset product shall be used as the final recommended product.
[0080] If the preset product is not a product in the initial recommended product combination and is not a pre-recommended product, then when the probability of the current user purchasing the preset product is greater than or equal to the second preset probability, the preset product will be used as the final recommended product.
[0081] The final recommended product portfolio consists of all the aforementioned final recommended products.
[0082] In one specific embodiment, the final recommended product combination is determined based on three pieces of information: the initial recommended product combination corresponding to the current user, the pre-recommended products, and the probability of the current user purchasing different preset products. If the first preset probability is 0.2 and the second preset probability is 0.5, the final recommended product table can be obtained, as shown in Table 3.
[0083] Table 3 Final Recommended Products List
[0084] Product Name Product type Purchase probability Is it recommended? Product A Pre-recommended products 0.20 yes Product B General products 0.49 no Product C General products 0.5 yes Product D Pre-recommended products 0.19 no Product E Initial Recommended Product Portfolio 0.20 yes Product F Initial Recommended Product Portfolio 0.19 no
[0085] If the preset product is a product in the initial recommended product combination or a pre-recommended product, then when the probability of the current user purchasing the preset product is greater than or equal to 0.2, the preset product is used as the final recommended product; if the preset product is neither a product in the initial recommended product combination nor a pre-recommended product, then when the probability of the current user purchasing the preset product is greater than or equal to 0.5, the preset product is used as the final recommended product. In Table 3, products A, C, and E are the final recommended products.
[0086] The product combination recommendation method provided in this invention obtains the current user's feature data and consumption data; based on the current user's feature data, it determines the customer group corresponding to the current user from a preset customer group; based on the customer group corresponding to the current user and the preset customer group, it determines the initial recommended product combination corresponding to the current user; it inputs the current user's feature data into a pre-trained machine learning model to obtain the probability of the current user purchasing different preset products; based on the initial recommended product combination corresponding to the current user and the probability of the current user purchasing different preset products, it determines the final recommended product combination; this improves the rationality of product combination recommendations, thereby improving the accuracy of product combination recommendations.
[0087] The product recommendation method provided in this invention integrates three types of information: the initial recommended product combination for the current user, the pre-recommended products, and the probability of the current user purchasing different preset products. This data is used to determine the final recommended product combination, significantly improving the rationality of the product recommendation while fully utilizing data resources. The method also considers the product's cyclical nature during the machine learning modeling process and the determination of pre-recommended products, thus ensuring the effectiveness of the product recommendation results.
[0088] The product recommendation method provided in this invention can be constructed based on artificial intelligence (AI). It acquires and processes relevant data using AI technology to achieve unattended, AI-powered product recommendation. AI is the theory, method, technology, and application system that uses digital computers or computers-controlled machines to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.
[0089] Foundational technologies for artificial intelligence generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interactive systems, and mechatronics. AI software technologies mainly encompass computer vision, robotics, biometrics, speech processing, natural language processing, and machine learning / deep learning.
[0090] Figure 2 This is a schematic diagram of the combined product recommendation device according to an embodiment of the present invention, as shown below. Figure 2 As shown, the combined product recommendation device 20 includes a data acquisition module 21, an initial recommendation module 22, a probability acquisition module 23, and a final recommendation module 24;
[0091] The data acquisition module 21 is used to acquire the current user's feature data and consumption data;
[0092] The initial recommendation module 22 is used to determine the customer group corresponding to the current user in a preset customer group based on the current user's feature data, and to determine the initial recommended product combination corresponding to the current user based on the customer group corresponding to the current user and the preset customer group. The preset customer group is determined based on the feature data and consumption data of a preset number of users. The preset customer group includes the customer group to which each user belongs in the preset number of users and the initial recommended product combination corresponding to each customer group.
[0093] The probability acquisition module 23 is used to input the feature data of the current user into a pre-trained machine learning model to obtain the probability of the current user purchasing different preset products. The machine learning module is trained based on sample feature data and sample consumption data.
[0094] The final recommendation module 24 is used to determine the final recommended product combination based on the initial recommended product combination corresponding to the current user and the probability of the current user purchasing different preset products.
[0095] In an optional implementation, the product recommendation device 20 further includes a preset customer group determination module, which is used to acquire feature data and consumption data of a preset number of users; to create user profile data by profiling the feature data; to determine all customer groups of the preset number of users and the customer group to which each user belongs based on the user profile data; and to determine the initial recommended product combination corresponding to each customer group based on the customer group to which each user belongs and the consumption data of each user.
[0096] In an optional implementation, the preset customer group determination module is further configured to determine customer group factors and representative attributes of the customer group factors based on the user profile data, and determine all customer groups of the preset number of users and the customer group to which each user in the preset number of users belongs based on the customer group factors and the representative attributes of the customer group factors.
[0097] In an optional implementation, the initial recommendation module 22 is further configured to determine the customer group factor corresponding to the current user and the representative attribute corresponding to the customer group factor based on the feature data of the current user, compare the similarity of the representative attribute corresponding to the customer group factor with the representative attribute of the customer group factor corresponding to all customer groups, and determine the customer group corresponding to the current user based on the similarity comparison result.
[0098] In an optional implementation, the final recommendation module 24 determines the final recommended product combination based on the initial recommended product combination corresponding to the current user and the probability of the current user purchasing different preset products, including:
[0099] If the preset product is a product in the initial recommended product combination, then when the probability of the current user purchasing the preset product is greater than or equal to the first preset probability, the preset product is used as the final recommended product.
[0100] If the preset product is not a product in the initial recommended product combination, then when the probability of the current user purchasing the preset product is greater than or equal to the second preset probability, the preset product will be used as the final recommended product.
[0101] The final recommended product portfolio consists of all the aforementioned final recommended products.
[0102] In an optional implementation, the current user's consumption data includes the products the current user has historically purchased and the time of purchase. The combined product recommendation device 20 further includes a pre-recommended product determination module, which is used to determine pre-recommended products based on the products the current user has historically purchased and the time of purchase. Correspondingly, the final recommendation module 24 is also used to determine the final recommended product combination based on the initial recommended product combination corresponding to the current user, the pre-recommended products, and the probability of the current user purchasing different preset products.
[0103] In an optional implementation, the final recommendation module 24 determines the final recommended product combination based on the initial recommended product combination corresponding to the current user, the pre-recommended products, and the probability of the current user purchasing different preset products, including:
[0104] If the preset product is a product in the initial recommended product combination, or is the pre-recommended product, then when the probability of the current user purchasing the preset product is greater than or equal to the first preset probability, the preset product shall be used as the final recommended product.
[0105] If the preset product is not a product in the initial recommended product combination and is not a pre-recommended product, then when the probability of the current user purchasing the preset product is greater than or equal to the second preset probability, the preset product will be used as the final recommended product.
[0106] The final recommended product portfolio consists of all the aforementioned final recommended products.
[0107] The product recommendation device 20 provided in this embodiment of the invention acquires the current user's feature data and consumption data through a data acquisition module 21; the initial recommendation module 22 determines the customer group corresponding to the current user from a preset customer group based on the current user's feature data, and determines the initial recommended product combination corresponding to the current user based on the customer group corresponding to the current user and the preset customer group; the probability acquisition module 23 inputs the current user's feature data into a pre-trained machine learning model to obtain the probability of the current user purchasing different preset products; and the final recommendation module 24 determines the final recommended product combination based on the initial recommended product combination corresponding to the current user and the probability of the current user purchasing different preset products. This improves the rationality of product recommendation and thus improves the accuracy of product recommendation.
[0108] Figure 3This is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention. Figure 3 As shown, the electronic device 30 includes a processor 31 and a memory 32 communicatively connected to the processor 31.
[0109] The memory 32 stores program instructions for implementing the combined product recommendation method of any of the above embodiments.
[0110] The processor 31 is used to execute program instructions stored in the memory 32 to make combined product recommendations.
[0111] The processor 31 can also be referred to as a CPU (Central Processing Unit). The processor 31 may be an integrated circuit chip with signal processing capabilities. The processor 31 can also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. A general-purpose processor can be a microprocessor or any conventional processor.
[0112] This invention provides a storage medium that stores program instructions capable of implementing all the methods described above. The storage medium can be non-volatile or volatile. These program instructions can be stored in the storage medium as a software product, including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks, or terminal devices such as computers, servers, mobile phones, and tablets.
[0113] In the several embodiments provided by this invention, it should be understood that the disclosed apparatus 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 in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.
[0114] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each module can exist physically separately, or two or more modules can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. The above are merely embodiments of the present invention and do not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.
[0115] The above description is merely an embodiment of the present invention. It should be noted that those skilled in the art can make improvements without departing from the inventive concept of the present invention, but these improvements all fall within the protection scope of the present invention.
Claims
1. A method for recommending combined products, characterized in that, Includes the following steps: Obtain the current user's characteristic data and consumption data; Based on the current user's characteristic data, determine the customer group corresponding to the current user in the preset customer group. Based on the customer group corresponding to the current user and the preset customer group, determine the initial recommended product combination corresponding to the current user. The preset customer group is determined based on the characteristic data and consumption data of a preset number of users. The preset customer group includes the customer group to which each user belongs in the preset number of users and the initial recommended product combination corresponding to each customer group. The current user's feature data is input into a pre-trained machine learning model to obtain the probability that the current user will purchase different preset products. The machine learning model is trained based on sample feature data and sample consumption data. A final recommended product combination is determined based on the initial recommended product combination for the current user and the probability of the current user purchasing different preset products. This determination includes: if the preset product is a product in the initial recommended product combination, then if the probability of the current user purchasing the preset product is greater than or equal to a first preset probability, the preset product is used as the final recommended product; if the preset product is not a product in the initial recommended product combination, then if the probability of the current user purchasing the preset product is greater than or equal to a second preset probability, the preset product is used as the final recommended product, wherein the second preset probability is greater than the first preset probability; and all of the final recommended products are used as the final recommended product combination.
2. The method for recommending combined products according to claim 1, characterized in that, The steps for determining the preset customer group include: Obtain feature data and consumption data of a preset number of users; The feature data is used to create a profile to obtain user profile data. Based on the user profile data, all customer groups of the preset number of users and the customer group to which each user in the preset number of users belongs are determined. Based on the customer group to which each user belongs and the consumption data of each user, an initial recommended product combination is determined for each customer group.
3. The method for recommending combined products according to claim 2, characterized in that, The step of determining all customer groups of the preset number of users and the customer group to which each user belongs, based on the user profile data, includes: Based on the user profile data, customer group factors and representative attributes of the customer group factors are determined. Based on the customer group factors and representative attributes of the customer group factors, all customer groups of the preset number of users and the customer group to which each user in the preset number of users belongs are determined.
4. The method for recommending combined products according to claim 3, characterized in that, Based on the current user's characteristic data, determine the customer group corresponding to the current user from the preset customer groups, including: Based on the current user's feature data, determine the customer group factor corresponding to the current user and the representative attribute corresponding to the customer group factor. Compare the similarity between the representative attribute corresponding to the customer group factor and the representative attribute corresponding to the customer group factor of all customer groups. Determine the customer group corresponding to the current user based on the similarity comparison result.
5. The method for recommending combined products according to claim 1, characterized in that, The current user's consumption data includes the products the current user has purchased in the past and the time of purchase. The method also includes determining pre-recommended products based on the current user's past purchases and the time of purchase. Accordingly, the final recommended product combination is determined based on the initial recommended product combination corresponding to the current user and the probability of the current user purchasing different preset products, including: determining the final recommended product combination based on the initial recommended product combination corresponding to the current user, the pre-recommended products, and the probability of the current user purchasing different preset products.
6. The method for recommending combined products according to claim 5, characterized in that, The final recommended product combination is determined based on the initial recommended product combination for the current user, the pre-recommended products, and the probability of the current user purchasing different preset products, including: If the preset product is a product in the initial recommended product combination, or is the pre-recommended product, then when the probability of the current user purchasing the preset product is greater than or equal to the first preset probability, the preset product shall be used as the final recommended product. If the preset product is not a product in the initial recommended product combination and is not a pre-recommended product, then when the probability of the current user purchasing the preset product is greater than or equal to the second preset probability, the preset product is used as the final recommended product, wherein the second preset probability is greater than the first preset probability. The final recommended product portfolio consists of all the aforementioned final recommended products.
7. A combined product recommendation device, characterized in that, It includes a data acquisition module, an initial recommendation module, a probability acquisition module, and a final recommendation module; The data acquisition module is used to acquire the current user's feature data and consumption data; The initial recommendation module is used to determine the customer group corresponding to the current user in a preset customer group based on the current user's feature data, and to determine the initial recommended product combination corresponding to the current user based on the customer group corresponding to the current user and the preset customer group. The preset customer group is determined based on the feature data and consumption data of a preset number of users. The preset customer group includes the customer group to which each user belongs in the preset number of users and the initial recommended product combination corresponding to each customer group. The probability acquisition module is used to input the current user's feature data into a pre-trained machine learning model to obtain the probability that the current user will purchase different preset products. The machine learning model is trained based on sample feature data and sample consumption data. The final recommendation module is used to determine a final recommended product combination based on the initial recommended product combination corresponding to the current user and the probability of the current user purchasing different preset products. This determination includes: if the preset product is a product in the initial recommended product combination, then if the probability of the current user purchasing the preset product is greater than or equal to a first preset probability, using the preset product as the final recommended product; if the preset product is not a product in the initial recommended product combination, then if the probability of the current user purchasing the preset product is greater than or equal to a second preset probability, using the preset product as the final recommended product, wherein the second preset probability is greater than the first preset probability; and using all of the final recommended products as the final recommended product combination.
8. An electronic device, comprising a memory and a processor, wherein the memory stores a computer program executable by the processor, characterized in that, When the processor executes the computer program, it implements the combined product recommendation method as described in any one of claims 1 to 6.
9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the combined product recommendation method as described in any one of claims 1 to 6.
Citation Information
Patent Citations
Commodity recommendation method, device, medium and equipment
CN113724042A
Commodity recommendation method and device and computer readable storage medium
CN114764732A