Method, system and medium for generating maternal and infant product information based on user information analysis

Abstract

The present invention discloses a method, system and medium for generating maternal and infant product information based on user information analysis; it relates to the technical field of maternal and infant product recommendation, and is aimed at the scenario in which vertical maternal and infant product merchants combine marketing through physical store sales and online sales. This solution improves the method on the basis of the existing complementary product recommendation technology, obtains target product information based on user behavior information, constructs a first complementary product set that is complementary to the target product function, and constructs a second complementary product set that is complementary to the target product subject in combination with user basic information; screens out the first complementary product list and the second complementary product list, generates maternal and infant product information and pushes it; generates a complementary product set not only from the dimension of complementary function of the target product, but also from the dimension of complementary subject of the target product, which can improve the accuracy of maternal and infant product recommendation and realize personalized product recommendation for maternal and infant users.

Description

Technical Field

[0001] The present invention relates to the technical field of maternal and infant product recommendation, and in particular to a method, system and medium for generating maternal and infant product information based on user information analysis. Background Art

[0002] With the rapid development of network technology, more and more businesses have begun to engage in e-commerce and sell supplies through the Internet. Merchants selling online can basically be divided into two categories. One category is to directly join the current major e-commerce platforms to sell goods, and recommend goods through the resources allocated by the major e-commerce platforms. However, this method has certain restrictions for merchants, which are easily restricted by e-commerce platforms and have a large cost burden for recommendation resources. The other category is for certain special types of goods, such as maternal and child supplies. Merchants are equipped with chain physical stores on the one hand, and have designed their own online shopping platforms on the other hand.

[0003] For vertical maternal and infant products merchants who combine physical store sales with online sales, the product recommendation methods currently used by merchants on their own online shopping platforms are similar to those of major e-commerce companies. When applied to the marketing of vertical maternal and infant products, there are problems of inaccurate recommendations and large volume, resulting in a lack of competitiveness, low visitor conversion efficiency, and serious loss of registered users. Summary of the invention

[0004] The technical problem to be solved by the present invention is: for vertical maternal and infant product merchants, in the scenario of combining physical store sales with online sales, in the process of physical store product recommendation, excessive reliance on the business ability of shopping guides who are familiar with product information and customer information leads to inaccurate recommendations and heavy workload; this solution is aimed at vertical maternal and infant product merchants in the scenario of combining physical store sales with online sales, and improves the method on the basis of the existing complementary product recommendation technology, designs a method for generating maternal and infant product information based on user information analysis, obtains target product information based on user behavior information, and constructs a first complementary product set that is complementary to the target product function with the target product information as the center, and at the same time constructs a second complementary product set that is complementary to the target product subject with the target product information as the center and combined with user basic information; a first complementary product list and a second complementary product list are screened out according to the first complementary product set and the second complementary product set, and finally generates maternal and infant product information and pushes it; not only generates a complementary product set from the dimension of complementary function of the target product, but also generates a complementary product set from the dimension of complementary subject of the target product, which can improve the recommendation accuracy of maternal and infant products and realize personalized product recommendation for maternal and infant users.

[0005] The present invention is achieved through the following technical solutions:

[0006] This solution provides a method for generating maternal and infant product information based on user information analysis, including:

[0007] Obtaining basic user information and user behavior information; the user behavior information includes direct behavior information or indirect behavior information;

[0008] Acquire target products based on the user behavior information, construct a first complementary product set with the target products as the center and complementary to the target products in terms of function, and construct a second complementary product set with the target products as the center and in combination with the basic information of the user and complementary to the target products in terms of stage and subject;

[0009] After respectively screening the first complementary product set and the second complementary product set for complementary correlation, a first complementary product list and a second complementary product list are obtained;

[0010] The maternal and infant product information is generated and pushed according to the first complementary product list and the second complementary product list.

[0011] Working principle of this scheme: For vertical maternal and infant product merchants in the scenario of combining marketing through physical store sales and online sales, the physical store product recommendation process relies too much on the business ability of the shopping guide who is familiar with the product information and customer information, resulting in inaccurate recommendations and heavy workload; and the product recommendation method adopted by the merchants' self-built online shopping platforms is similar to the recommendation methods of major e-commerce companies. When applied to the marketing process of vertical maternal and infant products, there are problems of inaccurate recommendations and large volume; resulting in lack of competitiveness, low visitor conversion efficiency, and serious loss of registered users; This scheme is aimed at vertical maternal and infant product merchants in the scenario of combining marketing through physical store sales and online sales, and designs a method for generating maternal and infant product information based on user information analysis. The maternal and infant product information generated in this scheme is mainly for complementary products of the target products. Since maternal and infant products have large stage differences, these products need to be adjusted according to the specific needs of infants at different growth stages, and at different The beneficiaries of maternal and infant products at different stages are also different (such as the postpartum period of a parturient overlaps with the neonatal stage and infant stage of a baby). The recommendation methods of traditional major e-commerce companies do not take into account the particularity of the large stage differences in maternal and infant products, so there are problems of inaccurate recommendations and large volume. This solution obtains target product information based on user behavior information, and constructs a first complementary product set that is complementary to the target product function with the target product information as the center. At the same time, a second complementary product set that is complementary to the target product subject is constructed with the target product information as the center and combined with the user's basic information. A first complementary product list and a second complementary product list are selected based on the first complementary product set and the second complementary product set, and finally maternal and infant product information is generated and pushed. Not only is a complementary product set generated from the dimension of complementary functions of the target products, but this solution also generates a complementary product set from the dimension of complementary beneficiaries of the target products, which can improve the accuracy of maternal and infant product recommendations and achieve personalized recommendations for maternal and infant users.

[0012] A further optimization scheme is that the method of obtaining target products based on the user behavior information includes:

[0013] For direct behavioral information:

[0014] All products involved in the user's direct behavior process are regarded as target products; in the specific implementation process, direct behavior information includes target products directly expressed in physical stores, products involved in completed online orders, and products added to shopping carts online.

[0015] Regarding indirect behavior information: During the specific implementation process, indirect behavior information mainly includes browsing information related to supplies, searching information related to supplies, and information related to participation in marketing activities.

[0016] Obtain all supplies and operation information of each supply involved in the indirect behavior information;

[0017] Calculate the expectation Q of each product based on the product operation information;

[0018] Sort the products from large to small according to the expected degree Q, and take the top n products in the expected degree Q as the target products.

[0019] A further optimization scheme is that the calculation of the expectation degree Q of each product based on the product operation information includes the following method:

[0020] Based on the operation information of the current product, the expected degree Q of the current product is calculated according to the following formula:

[0021]

[0022] Among them, S represents the number of searches for the current product; L represents the total intended views of the current product, which is the sum of the views of the product introduction page of the current product and the views of the positive review page of the current product; L1 represents the views of the current product introduction page; T1 represents the shortest viewing time of the current product introduction page; T2 represents the longest viewing time of the current product introduction page; T represents the total viewing time of the current product introduction page; t3 represents the shortest viewing time of the current product negative review page; t4 represents the longest viewing time of the current product negative review page; t represents the total viewing time of the current product negative review page; L2 represents the total views of the current product negative review page; m represents the activity tag participation parameter; m=1 represents that the user has participated in the activity, and m=0 represents that the user has not participated in the activity; k represents the correlation coefficient between the activity tag and the current product; ln represents the logarithmic function.

[0023] A further optimization scheme is that the first complementary product set is constructed with the target product as the center to complement the target product, including the following method:

[0024] Input each target product into a constructed functional complementary product neural network to obtain a first complementary product set; the functional complementary product neural network establishes a product graph according to the functional complementary relationship of the target products, and applies a graph attention network on the product graph to obtain a first feature of the target product, and then fuses the multimodal features of the target product into the first feature based on a common attention mechanism and a matrix decomposition bilinear pooling method, and uses the first feature after fusing the multimodal features of the target product to establish a complementary relationship between the products, and finally obtains a first complementary product set.

[0025] A further optimization scheme is that the second complementary product set is constructed with the target product as the center and combined with the basic information of the user, which is complementary to the target product in terms of stage and subject, including the method:

[0026] T1, taking infants as the beneficiary, the user's gestational period is first divided into multiple stages A, and taking pregnant women as the beneficiary, the user's gestational period is secondly divided into multiple stages B; wherein stage A and stage B include the same stage and overlapping stages;

[0027] T2, in the first division with infants as the beneficiary or the second division with pregnant women as the beneficiary, obtaining the current stage Ai or Bi of the user according to the basic information of the user;

[0028] T3, obtain the current target product and determine whether the current target product contains the attribute information of the stage. If so, proceed to step T4, otherwise, proceed to step T5;

[0029] T4, obtaining the stage x to which the current target product belongs according to the stage attribute information contained in the current target product, and determining whether the stage x is consistent with the stage Ai or Bi. If so, proceed to step T6; otherwise, proceed to step T5;

[0030] T5, using the product corresponding to the current target product in stage Ai or Bi as the second complementary product;

[0031] T6, pre-set the stage threshold to divide stage x into: early stage x and late stage x; if the user is currently in the early stage x, there is no second complementary product; if the user is in the late stage x, the product corresponding to the current target product in the next stage of stage x is used as the second complementary product;

[0032] T7, taking the set of second complementary products as the second complementary product set of the current target product.

[0033] A further optimization scheme is that step T4 also includes: obtaining the stage x to which the target product belongs according to the stage attribute information contained in the target product, and judging whether stage x belongs to a cross-overlapping stage. If so, generating one or more cross-overlapping stage products as second complementary products under the second division with pregnant women as the beneficiaries or the first division with infants as the beneficiaries; otherwise, judging whether stage x is consistent with stage Ai or Bi, and if so, entering step T6; otherwise, entering step T5.

[0034] A further optimization scheme is that the first complementary product set and the second complementary product set are screened for complementary association to obtain the first complementary product list and the second complementary product list, including the following method:

[0035] Calculate the complementary correlation coefficient S1 between each first complementary product in the first complementary product set and the target product, and calculate the complementary correlation coefficient S2 between each second complementary product in the second complementary product set and the target product;

[0036] Setting a first correlation coefficient threshold value Y1, screening out first complementary products with correlation coefficient S1 greater than the first correlation coefficient threshold value Y1, and arranging the first complementary products from large to small according to the correlation coefficient S1 to obtain a list of first complementary products;

[0037] A second correlation coefficient threshold value Y2 is set, second complementary products with correlation coefficient S2 greater than the second correlation coefficient threshold value Y2 are screened out, and the second complementary products are arranged from large to small according to the correlation coefficient S2 to obtain a list of second complementary products.

[0038] A further optimization scheme is that the method of generating and pushing maternal and infant product information according to the first complementary product list and the second complementary product list includes:

[0039] Search for recommendation instructions. If a functional complementarity recommendation instruction is received, the maternal and infant product information is generated and pushed in the order of the first complementary product list. If a stage complementary recommendation instruction is received, the maternal and infant product information is generated and pushed in the order of the second complementary product list. If no recommendation instruction is received, the first j first complementary products in the first complementary product list and the first j second complementary products in the second complementary product list are screened out, and the maternal and infant product information is generated and pushed in an alternating order.

[0040] The present solution also provides a system for generating maternal and infant product information based on user information analysis, which is used to implement the above-mentioned method for generating maternal and infant product information based on user information analysis; the system comprises:

[0041] A collection module, used to obtain basic user information and user behavior information; the user behavior information includes direct behavior information or indirect behavior information;

[0042] A construction module is used to obtain target products based on the user behavior information, construct a first complementary product set with the target products as the center and complementary to the target products in terms of function, and construct a second complementary product set with the target products as the center and in combination with the basic information of the user and complementary to the target products in terms of stage and subject;

[0043] A screening module, used for screening the first complementary product set and the second complementary product set for complementary association respectively to obtain a first complementary product list and a second complementary product list;

[0044] The output module is used to generate and push maternal and infant product information according to the first complementary product list and the second complementary product list.

[0045] The present solution also provides a computer-readable medium on which a computer program is stored. The computer program is executed by a processor to implement the above-mentioned method for generating maternal and child product information based on user information analysis.

[0046] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0047] 1. The present invention provides a method, system and medium for generating information on maternal and infant products based on user information analysis; for vertical maternal and infant product merchants who combine marketing through physical store sales and online sales, the present invention improves the method on the basis of the existing complementary product recommendation technology, and designs a method for generating information on maternal and infant products based on user information analysis, obtains target product information based on user behavior information, and constructs a first complementary product set that is complementary to the target product function with the target product information as the center, and at the same time constructs a second complementary product set that is complementary to the target product subject with the target product information as the center and in combination with the user's basic information; a first complementary product list and a second complementary product list are screened out from the first complementary product set and the second complementary product set, and finally generates maternal and infant product information and pushes it; the complementary product set is generated not only from the dimension of complementary function of the target product, but also from the dimension of complementary subject of the target product, which can improve the accuracy of maternal and infant product recommendation and realize personalized product recommendation for maternal and infant users.

[0048] 2. The present invention provides a method, system and medium for generating information on maternal and infant products based on user information analysis; a push mode is also set according to user needs. When the user only wants ordinary functional complementary products, a functional complementary recommendation instruction is sent, and the recommended products only include functional complementary products in the first complementary products list; when the user only wants stage or main body complementary products, a stage complementary recommendation instruction is sent, and the recommended products only include stage and main body complementary products in the second complementary products list; when the user does not send a recommendation instruction, the top-ranked products in the first complementary products list and the second complementary products list can be pushed alternately, providing convenience for users and shopping guides. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and should not be regarded as limiting the scope. For those of ordinary skill in the art, other relevant drawings can be obtained based on these drawings without creative work. In the drawings:

[0050] Figure 1 A flowchart of a method for generating maternal and infant product information based on user information analysis;

[0051] Figure 2 This is a schematic diagram of the structure of the maternal and infant product information generation system based on user information analysis. DETAILED DESCRIPTION

[0052] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments and drawings. The exemplary embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention.

[0053] For vertical maternal and infant product merchants who combine sales through physical stores and online sales, the product recommendation methods currently used by merchants to build their own online shopping platforms are similar to the recommendation methods of major e-commerce companies. When applied to the marketing process of vertical maternal and infant products, there are problems such as inaccurate recommendations and large volume, resulting in lack of competitiveness, low visitor conversion efficiency, and serious loss of registered users. In view of this, the present invention provides the following embodiments to solve the above technical problems.

[0054] Embodiment 1: This embodiment provides a method for generating maternal and infant product information based on user information analysis, such as Figure 1 As shown, including:

[0055] Step 1: Obtain user basic information and user behavior information; the user behavior information includes direct behavior information or indirect behavior information;

[0056] Step 2: based on the user behavior information, target products are obtained, a first complementary product set is constructed with the target products as the center, which is complementary to the target products in terms of function, and a second complementary product set is constructed with the target products as the center and in combination with the basic information of the user, which is complementary to the target products in terms of stage and subject;

[0057] The method of obtaining target products based on the user behavior information includes:

[0058] For direct behavioral information:

[0059] All products involved in the user's direct behavior process are taken as target products; in the specific implementation process, direct behavior information includes target product information directly expressed in physical stores, products involved in completed online orders, and products added to the online shopping cart. The product information involved in completed online orders and the product information added to the online shopping cart mainly include the information generated by the most recent operation.

[0060] Regarding indirect behavior information: In the specific implementation process, indirect behavior information mainly includes browsing product-related information, searching product-related information, and marketing activity participation-related information. Browsing product-related information, searching product-related information, and marketing activity participation-related information also mainly refer to the information generated by the most recent operation.

[0061] Obtain all supplies and operation information of each supply involved in the indirect behavior information;

[0062] Calculate the expectation Q of each product based on the product operation information;

[0063] Based on the operation information of the current product, the expected degree Q of the current product is calculated according to the following formula:

[0064]

[0065] Among them, S represents the number of searches for the current product; L represents the total intended views of the current product, which is the sum of the views of the product introduction page of the current product and the views of the positive review page of the current product; L1 represents the views of the current product introduction page; T1 represents the shortest viewing time of the current product introduction page; T2 represents the longest viewing time of the current product introduction page; T represents the total viewing time of the current product introduction page; t3 represents the shortest viewing time of the current product negative review page; t4 represents the longest viewing time of the current product negative review page; t represents the total viewing time of the current product negative review page; L2 represents the total views of the current product negative review page; m represents the activity tag participation parameter; m=1 represents that the user has participated in the activity, and m=0 represents that the user has not participated in the activity; k represents the correlation coefficient between the activity tag and the current product; ln represents the logarithmic function.

[0066] Sort the products from large to small according to the expected degree Q, and take the top n products in the expected degree Q as the target products.

[0067] The method of constructing a first complementary product set with the target product as the center and having complementary functions to the target product includes the following method:

[0068] Input each target product into a constructed functional complementary product neural network to obtain a first complementary product set; the functional complementary product neural network establishes a product graph according to the functional complementary relationship of the target products, and applies a graph attention network on the product graph to obtain a first feature of the target product, and then fuses the multimodal features of the target product into the first feature based on a common attention mechanism and a matrix decomposition bilinear pooling method, and uses the first feature after fusing the multimodal features of the target product to establish a complementary relationship between the products, and finally obtains a first complementary product set.

[0069] The method of taking the target product as the center and combining the basic information of the user to construct a second complementary product set that is complementary to the target product in terms of stage and subject, includes the following method:

[0070] T1, taking infants as the beneficiary, the user's gestational period is first divided into multiple stages A, and taking pregnant women as the beneficiary, the user's gestational period is secondly divided into multiple stages B; wherein stage A and stage B include the same stage and overlapping stages;

[0071] In the specific implementation process of this step, the user's pregnancy period is first divided into multiple stages A with infants as the beneficiary, which can be divided according to the conventional method, that is, stage A includes: early pregnancy, mid-pregnancy, late pregnancy, neonatal period, infant period and early childhood period. The user's pregnancy period is secondly divided into multiple stages B with pregnant women as the beneficiary, which can be divided according to the conventional method, that is, stage B includes: early pregnancy, mid-pregnancy, late pregnancy and postpartum period; the postpartum period in stage B overlaps with the neonatal period and infant period in stage A, and belongs to the cross-overlapping stage; the early pregnancy, mid-pregnancy and late pregnancy in stage A are completely consistent with the early pregnancy, mid-pregnancy and late pregnancy in stage B, and belong to the same stage.

[0072] The stage differences of maternity and baby products are indeed quite large, because these products need to be adjusted according to the specific needs of pregnant women and babies at different growth stages. The early stage of pregnancy is 1-3 months of pregnancy, and basic vitamins and minerals need to be supplemented, as well as special skin care products for pregnant women to cope with early skin changes; the second stage of pregnancy is 4-6 months of pregnancy, and it is necessary to wear comfortable maternity clothes and underwear to adapt to the gradual increase in the size of the abdomen. To ensure sleep, you may need to use a special maternity pillow to support the abdomen and waist and improve sleep quality. The late stage of pregnancy is 7-9 months of pregnancy, and it is necessary to prepare delivery-related products, such as maternity pads, chest protection covers, massage oils, soothing bath products, etc.; the newborn stage is 0-2 months after birth, and it is necessary to use particularly soft and non-irritating baby skin care products, such as baby shower gel, moisturizer, and use baby bottles and milk powder, and the safety and easy cleaning of the bottles need to be considered. The early stage of infants is 2-6 months after the birth of newborns, and it is necessary to introduce complementary foods, such as baby complementary food machines, baby tableware, and toys that stimulate sensory development, such as colorful toys and musical toys. The middle infant stage is from 6 to 12 months old, when a newborn needs to use a crawling mat and a walker; when the baby starts to feed himself, leak-proof cups and easy-to-grip tableware need to be prepared. The late infant stage is from 1 to 2 years old, when toys that can promote cognitive and language development, such as puzzles and building blocks, need to be selected.

[0073] T2. Under the first division with infants as the beneficiaries or the second division with pregnant women as the beneficiaries, the user's current stage Ai or Bi is obtained based on the user's basic information. In the specific implementation process, the user will fill in the corresponding pregnancy time information when registering an account, and the user's current stage can be calculated based on this information.

[0074] T3, obtain the current target product and determine whether the current target product contains stage attribute information. If so, enter step T4, otherwise, enter step T5; in the specific implementation process, target products such as milk powder stage I, milk powder stage II, and baby food machines contain stage attribute information. Based on the above target products, the stage to which the target product belongs can be directly inferred; while target products such as shower gel, moisturizer, and tableware do not contain stage attribute information.

[0075] T4, obtaining the stage x to which the current target product belongs according to the stage attribute information contained in the current target product, and determining whether the stage x is consistent with the stage Ai or Bi. If so, proceed to step T6; otherwise, proceed to step T5;

[0076] Step T4 also includes: obtaining the stage x to which the target product belongs according to the stage attribute information contained in the target product, and judging whether the stage x belongs to a cross-overlapping stage. If so, generating one or more cross-overlapping stage products as second complementary products under the second division with pregnant women as the beneficiaries or the first division with infants as the beneficiaries; otherwise, judging whether the stage x is consistent with the stage Ai or Bi, and if so, entering step T6; otherwise, entering step T5.

[0077] T5, using the product corresponding to the current target product in stage Ai or Bi as the second complementary product;

[0078] T6, pre-set stage thresholds to divide stage x into: early stage x and late stage x; if the user is currently in the early stage x, there is no second complementary product; if the user is in the late stage x, the products corresponding to the current target products in the next stage of stage x are used as the second complementary products; for the early pregnancy, mid-pregnancy, late pregnancy, neonatal period and early childhood, a threshold can be set to evenly divide the above stages into early stage and late stage. When the user is currently in the early stage, no complementary products may be generated; when the user is currently in the late stage, the complementary products of the next stage may be used as complementary products;

[0079] T7, taking the set of second complementary products as the second complementary product set of the current target product.

[0080] Conventional product recommendations based on functional complementarity, when directly applied to the marketing of maternal and infant products, may result in inaccurate recommendations and fail to accurately meet user needs (e.g., the target product is milk powder). This solution takes into account the stage differences in maternal and infant products. In the process of recommending complementary products, products with stage associations are also included in the list of complementary products to increase the selection range of recommended products. At the same time, this solution also takes into account the differences in the subjects of maternal and infant products. In the process of recommending complementary products, products with associations with the subjects of the products are also included in the list of complementary products to further increase the selection range of recommended products. (For example, in the traditional complementary product recommendation method, the target product is milk powder, and the recommended complementary products may include milk powder spoons, milk bottles, etc., while through the above method of this application, in addition to the above milk powder spoons and milk bottles, the recommended complementary products also include milk powder or complementary food for the next stage, as well as postpartum recovery products for mothers who are the beneficiaries.)

[0081] In this embodiment, the target product is milk powder stage I as an example. According to the above method, the second complementary product set is obtained, which corresponds to the newborn stage. When the user is currently in the infant stage, the second complementary product (stage complementary product) is milk powder stage II corresponding to the infant stage; when the user is currently in the early stage of the newborn stage, there is no second complementary product; when the user is currently in the late stage of the newborn stage, the second complementary product (stage complementary product) is milk powder stage II corresponding to the infant stage; and at this time, under the second division with pregnant women as the beneficiary, it belongs to the postpartum stage, and the second complementary product (subject complementary product) is maternity products.

[0082] Step 3: Screen the first complementary product set and the second complementary product set for complementary correlation to obtain a first complementary product list and a second complementary product list; during the implementation process, this step specifically includes the following method:

[0083] The complementary correlation coefficient S1 between each first complementary product in the first complementary product set and the target product is calculated, and the complementary correlation coefficient S2 between each second complementary product in the second complementary product set and the target product is calculated; in specific implementation, the complementary correlation coefficient S1 and the complementary correlation coefficient S2 can be calculated by using a correlation calculation method such as the Kendall τ correlation coefficient, the Spearman rank correlation coefficient, and the Jaccard similarity coefficient.

[0084] Setting a first correlation coefficient threshold value Y1, screening out first complementary products with correlation coefficient S1 greater than the first correlation coefficient threshold value Y1, and arranging the first complementary products from large to small according to the correlation coefficient S1 to obtain a list of first complementary products;

[0085] A second correlation coefficient threshold value Y2 is set, second complementary products with correlation coefficient S2 greater than the second correlation coefficient threshold value Y2 are screened out, and the second complementary products are arranged from large to small according to the correlation coefficient S2 to obtain a list of second complementary products.

[0086] Since the number of recommended products is limited, a list of complementary products can be constructed based on complementary correlation, and products with weak complementary correlation can be screened out. Only products with strong complementary correlation can be recommended. The order of recommendation can also be arranged according to the complementary correlation. The stronger the complementary correlation, the higher the recommendation order.

[0087] Step 4: Generate maternal and infant product information based on the first complementary product list and the second complementary product list and push it; this step specifically includes the following method:

[0088] Search for recommendation instructions. If a functional complementarity recommendation instruction is received, the information on maternal and infant products is generated and pushed in the order of the first complementary products list. If a stage complementary recommendation instruction is received, the information on maternal and infant products is generated and pushed in the order of the second complementary products list. If no recommendation instruction is received, the first j first complementary products in the first complementary products list and the first j second complementary products in the second complementary products list are screened out, and the information on maternal and infant products is generated and pushed in an alternating order. This solution also sets a push method according to the needs of the user. When the user only wants ordinary functional complementary products, a functional complementarity recommendation instruction is sent, and the recommended products only include functional complementary products in the first complementary products list; when the user only wants stage complementary products or main body complementary products, a stage complementary recommendation instruction is sent, and the recommended products only include stage and main body complementary products in the second complementary products list; when the user does not send a recommendation instruction, the top-ranked products in the first complementary products list and the second complementary products list can be pushed alternately.

[0089] Example 2

[0090] This embodiment provides a system for generating maternal and infant product information based on user information analysis, such as Figure 2 As shown, the method for generating maternal and infant product information based on user information analysis described in Example 1 is implemented; the system includes:

[0091] A collection module, used to obtain basic user information and user behavior information; the user behavior information includes direct behavior information or indirect behavior information;

[0092] A construction module is used to obtain target products based on the user behavior information, construct a first complementary product set with the target products as the center and complementary to the target products in terms of function, and construct a second complementary product set with the target products as the center and in combination with the basic information of the user and complementary to the target products in terms of stage and subject;

[0093] A screening module, used for screening the first complementary product set and the second complementary product set for complementary association respectively to obtain a first complementary product list and a second complementary product list;

[0094] The output module is used to generate and push maternal and infant product information according to the first complementary product list and the second complementary product list.

[0095] Example 3

[0096] This embodiment provides a computer-readable medium on which a computer program is stored. The computer program is executed by a processor to implement the method for generating maternal and infant product information based on user information analysis as described in Embodiment 1; specifically, the following steps are performed:

[0097] Step 1: Obtain user basic information and user behavior information; the user behavior information includes direct behavior information or indirect behavior information;

[0098] Step 2: based on the user behavior information, target products are obtained, a first complementary product set is constructed with the target products as the center, which is complementary to the target products in terms of function, and a second complementary product set is constructed with the target products as the center and in combination with the basic information of the user, which is complementary to the target products in terms of stage and subject;

[0099] Step 3: Screening the first complementary product set and the second complementary product set for complementary correlation respectively to obtain a first complementary product list and a second complementary product list;

[0100] Step 4: Generate maternal and infant product information based on the first complementary product list and the second complementary product list and push it.

[0101] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for generating maternal and infant product information based on user information analysis, characterized in that: include: Obtain basic user information and user behavior information; The user behavior information includes direct behavior information or indirect behavior information; Acquire target products based on the user behavior information, construct a first complementary product set with the target products as the center and complementary to the target products in terms of function, and construct a second complementary product set with the target products as the center and in combination with the basic information of the user and complementary to the target products in terms of stage and subject; Obtaining target products based on the user behavior information, including methods: For direct behavioral information: All products involved in the user's direct behavior process are taken as target products; For indirect behavioral information: Obtain all supplies and operation information of each supply involved in the indirect behavior information; Calculate the expectation Q of each product based on the product operation information; Sort the products from large to small according to the expected degree Q, and take the top n products in the expected degree Q as the target products; The method of taking the target product as the center and combining the basic information of the user to construct a second complementary product set that is complementary to the target product in terms of stage and subject, includes the following method: T1, taking infants as the beneficiary, the user's gestational period is first divided into multiple stages A, and taking pregnant women as the beneficiary, the user's gestational period is secondly divided into multiple stages B; wherein stage A and stage B include the same stage and overlapping stages; T2, in the first division with infants as the beneficiary or the second division with pregnant women as the beneficiary, obtaining the current stage Ai or Bi of the user according to the basic information of the user; T3, obtain the current target product and determine whether the current target product contains the attribute information of the stage. If so, proceed to step T4, otherwise, proceed to step T5; T4, obtaining the stage x to which the current target product belongs according to the stage attribute information contained in the current target product, and determining whether the stage x is consistent with the stage Ai or Bi. If so, proceeding to step T6; Otherwise, go to step T5; T5, using the product corresponding to the current target product in stage Ai or Bi as the second complementary product; T6, pre-setting the stage threshold to divide stage x into: early stage x and late stage x; If the user is currently in the early stage of stage x, there is no second complementary product; if the user is in the late stage x, the product corresponding to the current target product in the next stage of stage x is used as the second complementary product; T7, taking the set of second complementary products as the second complementary product set of the current target product; After respectively screening the first complementary product set and the second complementary product set for complementary correlation, a first complementary product list and a second complementary product list are obtained; The maternal and infant product information is generated and pushed according to the first complementary product list and the second complementary product list.

2. The method for generating maternal and infant product information based on user information analysis according to claim 1, characterized in that: The method of calculating the expectation degree Q of each product based on the product operation information includes: Based on the operation information of the current product, the expected degree Q of the current product is calculated according to the following formula: ； Among them, S represents the number of searches for the current product; L represents the total intended views of the current product, which is the sum of the views of the product introduction page of the current product and the views of the positive review page of the current product; L1 represents the views of the current product introduction page; T1 represents the shortest viewing time of the current product introduction page; T2 represents the longest viewing time of the current product introduction page; T represents the total viewing time of the current product introduction page; t3 represents the shortest viewing time of the current product negative review page; t4 represents the longest viewing time of the current product negative review page; t represents the total viewing time of the current product negative review page; L2 represents the total views of the current product negative review page; m represents the activity tag participation parameter; m=1 represents that the user has participated in the activity, and m=0 represents that the user has not participated in the activity; k represents the correlation coefficient between the activity tag and the current product; ln represents the logarithmic function.

3. The method for generating maternal and infant product information based on user information analysis according to claim 1, characterized in that: The method of constructing a first complementary product set with the target product as the center and having complementary functions to the target product includes the following method: Input each target product into a constructed functional complementary product neural network to obtain a first complementary product set; the functional complementary product neural network establishes a product graph according to the functional complementary relationship of the target products, and applies a graph attention network on the product graph to obtain a first feature of the target product, and then fuses the multimodal features of the target product into the first feature based on a common attention mechanism and a matrix decomposition bilinear pooling method, and uses the first feature after fusing the multimodal features of the target product to establish a complementary relationship between the products, and finally obtains a first complementary product set.

4. The method for generating maternal and infant product information based on user information analysis according to claim 1, characterized in that: Step T4 also includes: obtaining the stage x to which the target product belongs according to the stage attribute information contained in the target product, and judging whether the stage x belongs to a cross-overlapping stage. If so, generating one or more cross-overlapping stage products as second complementary products under the second division with pregnant women as the beneficiaries or the first division with infants as the beneficiaries; otherwise, judging whether the stage x is consistent with the stage Ai or Bi, and if so, entering step T6; otherwise, entering step T5.

5. The method for generating maternal and infant product information based on user information analysis according to claim 1, characterized in that: The method of obtaining the first complementary product list and the second complementary product list by respectively screening the first complementary product set and the second complementary product set for complementary association degrees comprises the following steps: Calculate the complementary correlation coefficient S1 between each first complementary product in the first complementary product set and the target product, and calculate the complementary correlation coefficient S2 between each second complementary product in the second complementary product set and the target product; Setting a first correlation coefficient threshold value Y1, screening out first complementary products with correlation coefficient S1 greater than the first correlation coefficient threshold value Y1, and arranging the first complementary products from large to small according to the correlation coefficient S1 to obtain a list of first complementary products; A second correlation coefficient threshold value Y2 is set, second complementary products with correlation coefficient S2 greater than the second correlation coefficient threshold value Y2 are screened out, and the second complementary products are arranged from large to small according to the correlation coefficient S2 to obtain a list of second complementary products.

6. The method for generating maternal and infant product information based on user information analysis according to claim 5, characterized in that: The method of generating and pushing maternal and infant product information according to the first complementary product list and the second complementary product list includes: Search for recommendation instructions. If a functional complementarity recommendation instruction is received, the maternal and infant product information is generated and pushed in the order of the first complementary product list. If a stage complementary recommendation instruction is received, the maternal and infant product information is generated and pushed in the order of the second complementary product list. If no recommendation instruction is received, the first j first complementary products in the first complementary product list and the first j second complementary products in the second complementary product list are screened out, and the maternal and infant product information is generated and pushed in an alternating order.

7. A maternal and infant product information generation system based on user information analysis, characterized in that: A method for generating maternal and infant product information based on user information analysis according to any one of claims 1 to 6; the system comprises: A collection module, used to obtain basic user information and user behavior information; the user behavior information includes direct behavior information or indirect behavior information; A construction module is used to obtain target products based on the user behavior information, construct a first complementary product set with the target products as the center and complementary to the target products in terms of function, and construct a second complementary product set with the target products as the center and in combination with the basic information of the user and complementary to the target products in terms of stage and subject; A screening module, used for screening the first complementary product set and the second complementary product set for complementary association respectively to obtain a first complementary product list and a second complementary product list; The output module is used to generate and push maternal and infant product information according to the first complementary product list and the second complementary product list.

8. A computer readable medium having a computer program stored thereon, characterized in that: The computer program is executed by a processor to implement the method for generating maternal and infant product information based on user information analysis as described in any one of claims 1 to 6.

Images