Product information recommendation method, device, equipment and storage medium

Abstract

The present application provides a method, device, equipment and storage medium for recommending product information, wherein the method includes: obtaining recommendation data from a recommendation platform, recalling products according to the recommendation data to obtain a recall set, determining at least one product to be recommended from the recall set according to preset rules to obtain a recommendation set, and sending recommendation information to the recommendation platform according to the recommendation set. In this technical solution, different product recall sets are recalled through recommendation data from different recommendation platforms, and then a recommendation set is determined from the product recall set, and recommendation information is sent to the recommendation platform to instruct the recommendation platform to recommend products, so that products can be accurately recommended to the user groups targeted by the recommendation platform, so that products can be accurately obtained by the required users, and the efficiency of product recommendation is improved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method, device, equipment and storage medium for recommending product information. Background Art

[0002] With the rapid development of Internet technology, the way users obtain the products they need has also changed from the traditional offline mode to the online mode. The online mode has a larger potential market. E-commerce can publish product information on various online website information platforms to spread it to users of various website information platforms, allowing more users to understand product information and ultimately enable users to obtain the products they need.

[0003] The existing cost-per-sale (CPS) is a common and effective form of product information dissemination. Product information is pushed to the user group targeted by the website information platform through the website information platform. When the user determines that he needs the product, he can enter the corresponding e-commerce platform through the website information platform to obtain the corresponding product. In the common product information recommendation process, an e-commerce company can recommend product information through multiple different websites at the same time.

[0004] However, the existing CPS information dissemination model is prone to push unnecessary product information to the user groups targeted by different website information platforms due to the different user groups targeted by each website information platform and the different needs of different user groups when recommending products to users. This causes users to be unable to obtain the information about the products they need in a timely manner, and the product recommendation efficiency is low. Summary of the invention

[0005] The present application provides a product information recommendation method, device, equipment and storage medium, which are used to solve the problem of low efficiency of existing product recommendations.

[0006] In a first aspect, an embodiment of the present application provides a method for recommending product information, including:

[0007] Acquire recommendation data from a recommendation platform, where the recommendation data is used to indicate recommendation effects of different products;

[0008] Performing product recall according to the recommendation data to obtain a recall set, wherein the recall set includes identifications and product information of multiple products, and the product information of each product includes value information at different value attributes, a recommendation result indicator, and at least one of an acquisition method;

[0009] According to a preset rule, at least one product to be recommended is determined from the recalled set to obtain a recommended set, wherein the recommended set includes an identifier and product information of the at least one product to be recommended;

[0010] According to the recommendation set, recommendation information is sent to the recommendation platform, where the recommendation information is used to instruct the recommendation platform to recommend the at least one product to be recommended.

[0011] In a possible design of the first aspect, recalling the product according to the recommendation data to obtain a recall set includes:

[0012] Performing a timed asynchronous recall of products according to preset timed tasks and the recommendation data to obtain a set of product candidates;

[0013] The product candidate set is filtered according to preset business rules to obtain the recalled set, wherein the preset business rules include at least one of a recommended fee rule and a product discount rule.

[0014] In another possible design of the first party, the preset rule includes a default sequence, and determining at least one product to be recommended from the recalled set according to the preset rule to obtain a recommended set includes:

[0015] Storing the recalled set in a preset database;

[0016] A default sequence of products in a recalled set stored in the preset database is obtained, and at least one product to be recommended is determined according to the default sequence to obtain a recommended set.

[0017] In another possible design of the first aspect, the preset rule includes a preset sorting rule, and determining at least one product to be recommended from the recalled set according to the preset rule to obtain the recommended set includes:

[0018] According to the preset sorting rule, the products in the recall set are sorted to obtain a product arrangement order;

[0019] At least one product to be recommended is determined according to the product arrangement order to obtain a recommendation set.

[0020] In another possible design of the first aspect, the preset sorting rule includes a first sorting rule and a second sorting rule, and sorting the products in the recall set according to the preset sorting rule to obtain a product arrangement order includes:

[0021] Sorting the products according to the first sorting rule and the product information of the products in the recalled set to obtain a product arrangement order, wherein the first sorting rule includes a preset sorting algorithm; or,

[0022] The products are sorted according to the second sorting rule and the product information of the product recall set to obtain a product arrangement order, wherein the second sorting rule includes at least one of a product price sorting rule, a product recommendation fee sorting rule and a product evaluation information sorting rule.

[0023] In another possible design of the first aspect, the step of performing product sorting according to the first sorting rule and product information of the products in the recalled set to obtain a product arrangement order includes:

[0024] Obtaining user tags, wherein the user tags include user attribute tags and user behavior tags;

[0025] The user tags and the product information of the products in the recall set are input into the preset sorting algorithm to sort the products to obtain a product arrangement order. The preset sorting algorithm includes a gradient boosting decision tree algorithm and a logistic regression algorithm.

[0026] In yet another possible design of the first aspect, obtaining the user tag includes:

[0027] Obtain user attribute data and user behavior data through online logs, real-time messages, and offline calculations;

[0028] Encoding the user attribute data to obtain a user attribute tag;

[0029] Discretizing the user behavior data to obtain a user behavior label;

[0030] The user attribute tag and the user behavior tag are stored in a preset database, and the preset database is read to obtain the user tag.

[0031] In another possible design of the first aspect, before inputting the user tag and the product information of the products in the recalled set into the preset sorting algorithm to sort the products and obtain the product arrangement order, the method further includes:

[0032] Performing algorithm configuration on the preset sorting algorithm according to the preset algorithm configuration;

[0033] Testing the preset sorting algorithm after algorithm configuration to obtain test data;

[0034] The preset sorting algorithm is configured and optimized according to the test data.

[0035] In another possible design of the first aspect, after determining at least one product to be recommended according to the product arrangement order and obtaining the recommendation set, the method further includes:

[0036] Obtain identification parameters, and cache the recommendation set to a preset database according to the identification parameters, wherein the identification parameters include a scene identification, a user identification, and test data.

[0037] In yet another possible design of the first aspect, sending the recommendation information to the recommendation platform according to the recommendation set includes:

[0038] Obtain the landing page of each product in the recommendation set;

[0039] Performing link conversion on the landing page of each product in the recommendation set to obtain a converted recommendation link, wherein the recommendation link is used to collect recommendation data of the recommendation platform;

[0040] The recommendation information is obtained according to the recommendation link and the product information of the product, and the recommendation information is sent to the recommendation platform.

[0041] In a second aspect, an embodiment of the present application provides a device for recommending product information, including:

[0042] A data acquisition module, used to acquire recommendation data from the recommendation platform, wherein the recommendation data is used to indicate the recommendation effects of different products;

[0043] A recall module, configured to recall products according to the recommendation data to obtain a recall set, wherein the recall set includes identifiers and product information of multiple products, and the product information of each product includes value information at different value attributes, a recommendation result indicator, and at least one of an acquisition method;

[0044] A determination module, configured to determine at least one product to be recommended from the recalled set according to a preset rule, and obtain a recommendation set, wherein the recommendation set includes an identification and product information of the at least one product to be recommended;

[0045] The recommendation module is used to send recommendation information to the recommendation platform according to the recommendation set, wherein the recommendation information is used to instruct the recommendation platform to recommend the at least one product to be recommended.

[0046] In a third aspect, an embodiment of the present application provides a device for recommending product information, including a memory and at least one processor;

[0047] The memory stores computer-executable instructions;

[0048] The at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor performs any of the above methods.

[0049] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, wherein the computer-readable storage medium has computer instructions stored therein, and the computer instructions are used to implement the method described above when executed by a processor.

[0050] The product information recommendation method, apparatus, device and storage medium provided in the embodiments of the present application use recommendation data from different recommendation platforms to recall different product recall sets, and then determine a recommendation set from the product recall sets, and send recommendation information to the recommendation platform to instruct the recommendation platform to recommend products. This enables products to be accurately recommended to the user groups targeted by the recommendation platform, so that products can be accurately acquired by the required users, thereby improving product recommendation efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 A schematic diagram of a scenario of a method for recommending product information provided in an embodiment of the present application;

[0052] Figure 2 A flowchart of a method for recommending product information provided in an embodiment of the present application;

[0053] Figure 3 A flowchart of a second embodiment of a method for recommending product information provided in an embodiment of the present application;

[0054] Figure 4 A flowchart of a third embodiment of a method for recommending product information provided in an embodiment of the present application;

[0055] Figure 5 A flowchart of a fourth embodiment of a method for recommending product information provided in an embodiment of the present application;

[0056] Figure 6 A flowchart of a fifth embodiment of a method for recommending product information provided in an embodiment of the present application;

[0057] Figure 7 A flowchart of Example 6 of a method for recommending product information provided in an embodiment of the present application;

[0058] Figure 8 A schematic diagram of the structure of a first embodiment of a device for pushing product information provided in an embodiment of the present application;

[0059] Fig. 9 A schematic diagram of the structure of a second embodiment of a device for recommending product information provided in an embodiment of the present application;

[0060] Fig.10 A schematic diagram of the structure of a recommended device for providing product information in an embodiment of the present application. DETAILED DESCRIPTION

[0061] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0062] Figure 1 A schematic diagram of a scenario of a method for recommending product information provided in an embodiment of the present application, such as Figure 1 As shown, the application scenario of this embodiment may be an express cabinet, which includes at least a display device 13 for displaying information and a container 14 for storing express delivery. In actual use, the user can input a pickup code on the interface of the display device 13. When the express cabinet recognizes the pickup code, it controls the container 14 to open and displays the location of the opened container 14 on the interface. The user then takes the express delivery out of the container 14.

[0063] In the process of picking up the express, there is information interaction between the user and the display device 13, and some information can be displayed on the interface of the display device 13 to recommend to the user, such as recent weather information, product information and recent news information.

[0064] Exemplarily, some product information is displayed on the interface of the display device 13 to recommend it to the user. If the user has a demand for the product after reading the product information, he or she can continue to click on the interface, and the interface will display the purchase method of the product (for example, scanning a QR code). The user can purchase the product by scanning the QR code. The product information displayed on the interface of the display device 13 can be downloaded from the server 12. Furthermore, the product provider can also determine the product to be recommended through the remote terminal 11, edit the content of the product information, and upload it to the server 12.

[0065] The above-mentioned express lockers can be used as an example of a recommendation platform. There are various recommendation platforms in daily life. These recommendation platforms can be used to recommend product information to their target users. However, in actual application, different recommendation platforms target different user groups, and the needs of different user groups are also different. This makes it easy for the recommendation platform to recommend product information that some users do not need, thereby reducing the efficiency of product recommendations.

[0066] In response to the above problems, the embodiments of the present application provide a method, apparatus, device and storage medium for recommending product information. The inventive concept is as follows: when recalling products and product information, in the same scenario, based on the user groups targeted by different recommendation platforms, different product recall sets are recalled for each different recommendation platform, thereby realizing recall sets that are tailored to each individual. This enables the recommendation platform to accurately recommend products to the required users when recommending products, thereby achieving a more accurate recommendation effect and improving recommendation efficiency.

[0067] The technical solution of the present application is described in detail below through specific embodiments. It should be noted that the following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0068] Figure 2 This is a flow chart of a method for recommending product information provided in an embodiment of the present application. This method can be applied to the above-mentioned remote terminal or device with information processing function. Figure 2 As shown, the method for recommending product information may include the following steps:

[0069] S201. Obtain recommendation data from a recommendation platform.

[0070] Among them, the recommendation data is used to indicate the recommendation effects of different products. For example, when the recommendation platform recommends products, some products have high demand from users, while some products have low demand from users. The user's click rate can be used to determine which products have high demand and which products have low demand, thereby determining the recommendation effects of different products.

[0071] In the embodiments of the present application, the recommendation platform can be a variety of different types of recommendation platforms. For example, physical devices such as express lockers and smart terminals that can interact with users can be used as recommendation platforms. Furthermore, the recommendation platform can also be some virtual websites, and users and websites also have information interaction. It can be understood that the user group targeted by express lockers is generally users who use online shopping more, while the user group targeted by smart terminals is mainly teenagers who use smart terminals more frequently.

[0072] For example, products can be classified according to categories, which may include electronic products, wearable apparel products, outdoor sports products, etc. Different categories can be further refined. For example, electronic products include mobile phones, computers, and household appliances, etc. These products are provided by product providers.

[0073] S202: recall products according to the recommended data to obtain a recall set.

[0074] The recall set includes identifications and product information of multiple products, and the product information of each product includes value information at different value attributes, recommendation result indicators, and at least one of acquisition methods.

[0075] Exemplarily, the product identification can be a digital code, such as SKU ID (Stock keeping Unit); the value information at different value attributes can be the lowest price of the product within a certain period of time and the optimal coupon discount ratio corresponding to the product; the recommendation result indicator can be the purchase volume achieved when the product is recommended through the recommendation platform, including 1-day order volume, 7-day order volume, 15-day order volume and 30-day order volume; the acquisition method can refer to whether the product is a group purchase product.

[0076] For example, when there is only one recommendation platform, among the products recommended by the recommendation platform, some products have better recommendation effects, then the products with better recommendation effects are mainly recalled to obtain a recall set. When there are multiple recommendation platforms, each recommendation platform will correspond to a recall set, thus obtaining a recall set with thousands of faces.

[0077] S203: According to preset rules, at least one product to be recommended is determined from the recalled set to obtain a recommended set.

[0078] The recommendation set includes at least one identifier and product information of a product to be recommended. For example, the identifier of the product to be recommended may be a SKU ID, and the product information of the product to be recommended includes the price of the product, whether there is a coupon, and the like.

[0079] In an embodiment of the present application, the preset rule may be a default sequence of products in a recall set. Specifically, when products are recalled to form a recall set, each product will be arranged in order of recall to form a default sequence, and the first N products of the default sequence will be determined from the recall set to obtain a recommended set, where N is a positive integer not less than 1.

[0080] Exemplarily, the preset rule may also be a preset sorting rule, which may be a related personalized sorting algorithm, such as a gradient boosting decision tree (GBDT). Specifically, the user's preferred products are determined through the gradient boosting decision tree sorting algorithm, and then the products to be recommended are determined based on the user's preferred products to obtain a recommendation set.

[0081] Exemplarily, the recall set can be stored in a preset database, such as a remote dictionary service (Redis, RemoteDictionary Server). When the preset sorting rule is loaded in the remote terminal and the preset sorting rule is available, the recommended set is obtained according to the preset sorting rule. When the preset sorting rule is not loaded in the remote terminal or the preset sorting rule is in an abnormal and unavailable state, the recall set is retrieved from the preset database and the recommended set is obtained according to the default sequence.

[0082] S204: Send recommendation information to the recommendation platform according to the recommendation set.

[0083] The recommendation information is used to instruct the recommendation platform to recommend at least one product to be recommended. Exemplarily, the recommendation information includes a purchase link, price information, and whether the product to be recommended has a coupon.

[0084] Exemplarily, the recommendation platform may be operated by a third-party promoter. Specifically, in order to maximize the recommendation profit, the third-party promoter may determine whether the product to be recommended needs to be displayed on the recommendation platform based on the recommendation information of the product to be recommended. For example, when the third-party promoter reads the recommendation information and finds that the product to be recommended does not have a coupon, the product to be recommended may not be displayed on the recommendation platform, and the user cannot view the product information of the product to be recommended and cannot purchase the product to be recommended.

[0085] Exemplarily, the remote terminal may communicate directly with the recommendation platform. When the recommendation platform sends a recommendation request to the remote terminal, the remote terminal sends recommendation information to the recommendation platform according to the recommendation request.

[0086] The embodiment of the present application performs product recall based on the recommendation data of different recommendation platforms to obtain different recall sets, and then determines the products to be recommended from the recall sets to obtain a recommendation set, so that the recommendation platform can recommend products required by the user group when making product recommendations, thereby achieving more accurate recommendation effects and improving recommendation efficiency.

[0087] Exemplarily, based on the above embodiments, Figure 3 The flowchart of the second embodiment of the method for recommending product information provided in the embodiment of the present application is as follows: Figure 3 As shown, the above step S202 can be specifically implemented by the following steps:

[0088] S301. Perform a scheduled asynchronous recall of products according to preset scheduled tasks and recommendation data to obtain a set of product candidates.

[0089] Exemplarily, the time interval of the scheduled task is configurable, for example, the time interval is 10 minutes, and when the scheduled task arrives, an asynchronous recall of the product is performed.

[0090] In an embodiment of the present application, the product candidate set includes identifications and product information of multiple products. Exemplarily, the product information includes the lowest price of the product, the best coupon discount ratio, the number of orders generated on the recommendation platform in 1 day, the number of orders generated on the recommendation platform in 7 days, the number of orders generated on the recommendation platform in 15 days, the number of orders generated on the recommendation platform in 30 days, whether the product is a group purchase, and other information.

[0091] S302: Filter the product candidate set according to preset business rules to obtain a recall set.

[0092] The preset business rules include at least one of a recommendation fee rule and a product discount rule. Specifically, the recommendation fee rule refers to the fee that the recommendation platform should earn when recommending a product to a user if the user purchases the product, and the product discount rule refers to whether there is a discount coupon when the user purchases a product.

[0093] Exemplarily, when products in the product candidate set fail to meet the preset business rules, these products are filtered out, and finally products meeting the preset business rules are obtained, and the identifications and product information of these products are collected in the recall set.

[0094] For example, when a recall set is obtained after filtering the product candidate set, the products in the recall set will have a default sequence, and the recommendation set can be obtained through the default sequence.

[0095] The embodiment of the present application performs a scheduled asynchronous recall of products and obtains a recall set by filtering through preset business rules, thereby filtering out some products that do not meet the requirements, reducing noise and improving product recommendation efficiency.

[0096] Optionally, in some embodiments, the above-mentioned preset sorting rules include a first sorting rule and a second sorting rule. Exemplarily, the first sorting rule may be a preset sorting algorithm, and the second sorting rule may be at least one of a product price sorting rule, a product recommendation fee sorting rule, and a product evaluation information sorting rule. Specifically, when the preset sorting rule is the first sorting rule, the products are sorted according to the first sorting rule and the product information of the recall set to obtain a product arrangement order. When the preset sorting rule is the second sorting rule, the products are sorted according to the second sorting rule and the product information of the recall set to obtain a product arrangement order. According to the product sorting order, the top N products are selected as the products to be recommended to obtain a recommendation set, where N is a positive integer not less than 1.

[0097] The embodiment of the present application divides the preset sorting rules into two different sorting rules, allowing operators to flexibly select appropriate sorting rules in different scenarios, sort the recall set, obtain the product arrangement order, and determine the recommended set based on the product arrangement order.

[0098] Exemplarily, based on the above embodiments, Figure 4 The flowchart of the third embodiment of the method for recommending product information provided in the embodiment of the present application is as follows: Figure 4 As shown, the above step of "sorting the products according to the first sorting rule and the product information of the recalled set to obtain the product arrangement order" can be specifically implemented by the following steps:

[0099] S401: Obtain user tags.

[0100] Among them, user tags include user attribute tags and user behavior tags. For example, user attribute tags include user's own attributes such as age, gender, occupation, education level, etc., and user behavior tags include user browsing, clicking, following, adding to cart, and purchasing, etc. Furthermore, user behavior tags can be user's behavior in the past 30 days.

[0101] S402: Input the user tags and the product information of the products in the recall set into a preset sorting algorithm to sort the products and obtain a product arrangement order.

[0102] Among them, the preset sorting algorithms include gradient boosting decision tree algorithm and logistic regression algorithm (LR, logistic regression).

[0103] The embodiment of the present application obtains user tags and sorts products through a gradient boosting decision tree algorithm and a logistic regression algorithm, and can obtain a product arrangement order according to the user's product preference sorting. Subsequently, personalized product recommendations can be made for different users, thereby improving the efficiency of product recommendations.

[0104] Exemplarily, based on the above embodiments, Figure 5 The flowchart of the fourth embodiment of the method for recommending product information provided in the embodiment of the present application is as follows: Figure 5 As shown, the above step S401 can be specifically implemented by the following steps:

[0105] S501. Obtain user attribute data and user behavior data through online logs, real-time messages and offline calculations.

[0106] S502: Encode the user attribute data to obtain a user attribute tag.

[0107] S503: discretize the user behavior data to obtain a user behavior label.

[0108] S504: Store the user attribute tag and the user behavior tag in a preset database, and read the preset database to obtain the user tag.

[0109] Exemplarily, when a user interacts with the recommendation platform, corresponding online logs, real-time messages and offline data will be generated. The online log may be the user's behavior log for the past two days, the real-time message may be the user's real-time behavior, and the offline data may be the user's behavior data for the past month and the user's own attribute information. User behavior data includes the user's real-time behavior data and the user's historical behavior data. Exemplarily, user behavior includes the user's browsing, clicking, following, adding to cart, and purchasing behaviors.

[0110] Optionally, a discretized user attribute label can be obtained by one-hot encoding the user attribute data, and the user behavior label can be obtained by discretizing the user behavior data according to the time dimension, for example, discretizing the user behavior data at four time points of 1 day, 3 days, 7 days and 30 days.

[0111] Exemplarily, the preset database may be a redis database.

[0112] The embodiment of the present application discretizes and encodes user attribute data and user behavior data to obtain user attribute labels and user behavior labels, and stores them in a preset database. When user labels are needed, it is only necessary to retrieve user attribute labels and user behavior labels from the preset database, thereby improving the sorting efficiency when the preset sorting algorithm is used for sorting.

[0113] Exemplarily, based on the above embodiments, Figure 6 The flowchart of the fifth embodiment of the method for recommending product information provided in the embodiment of the present application is as follows: Figure 6 As shown, the above step S402 further includes the following steps:

[0114] S601: Configure a preset sorting algorithm according to a preset algorithm configuration.

[0115] S602: Testing the preset sorting algorithm after algorithm configuration to obtain test data.

[0116] S603: Optimize the configuration of the preset sorting algorithm according to the test data.

[0117] In an embodiment of the present application, the preset sorting algorithm includes a variety of different algorithms. For example, an AB test method can be used to test the preset sorting algorithm after the algorithm configuration to obtain test data, wherein the test data includes traffic data. For example, the preset sorting algorithm using A configuration occupies 70% of the traffic, while the preset sorting algorithm using B configuration occupies 30% of the traffic. Furthermore, by configuring different algorithms for the preset sorting algorithm, the final product arrangement order will also be different.

[0118] Exemplarily, the preset sorting algorithms include a gradient boosting decision tree algorithm and a logistic regression algorithm.

[0119] By optimizing the configuration of a preset sorting algorithm, the embodiment of the present application can sort products using the optimized preset sorting algorithm to obtain a product arrangement order that better meets user preferences, achieve personalized product recommendations, and improve product recommendation efficiency.

[0120] Optionally, in some embodiments, after the recommendation set is obtained, the recommendation set may be cached in a preset database according to the representation parameters.

[0121] Among them, the identification parameters include scene identification, user identification and test data. Exemplarily, the scene identification can be a discount scene identification, for example, the scene identification of a certain product is 9.9 yuan instant purchase, and the test data can be the granularity of the AB experimental parameters obtained by using the above-mentioned AB testing method. The preset database can be a redis database.

[0122] In the embodiment of the present application, by caching the recommendation set in a preset database, recommendation information can be directly obtained according to the recommendation set in the preset database, which can improve the performance and stability of the recommendation service of the recommendation platform.

[0123] Exemplarily, based on the above embodiments, Figure 7 The flowchart of the sixth embodiment of the method for recommending product information provided in the embodiment of the present application is as follows: Figure 7 As shown, the above step S204 can be specifically implemented by the following steps:

[0124] S701. Obtain the landing page of each product in the recommended set.

[0125] In an embodiment of the present application, the landing page of the product refers to the web page displayed to the user after the user clicks or scans the QR code. The user can see the price, size and other related information of the product on this page, and can purchase the product by operating on this page.

[0126] S702: Perform link conversion on the landing page of each product in the recommendation set to obtain a converted recommendation link.

[0127] Among them, the recommendation link is used to collect recommendation data of the recommendation platform. In this embodiment, the recommendation link is used to record some behavioral data of users from clicking to placing an order to purchase a product, which may specifically include product exposure rate and user behavior data such as product clicks and order purchases.

[0128] For example, when the user clicks on the recommendation link and jumps to the order purchase interface through the recommendation link, the user's click behavior can be collected by the recommendation link record in this process as the recommendation data of the recommendation platform. Subsequently, the remote terminal or information processing device can perform product recall based on the recommendation data collected by the recommendation link to obtain a recall set, forming a closed-loop product recommendation optimization process.

[0129] It is understandable that relevant algorithms or computer programs can be embedded in the recommendation links to collect user behavior data such as product exposure and product clicks and order purchases. The specific implementation process of the algorithms or computer programs will not be elaborated here.

[0130] S703: Obtain recommendation information according to the recommendation link and the product information of the product, and send the recommendation information to the recommendation platform.

[0131] Exemplarily, the product information of a product includes whether there is a discount coupon for the product. After reading the recommendation information, the recommendation platform can determine whether to recommend the product to the user based on whether there is a discount coupon for the product.

[0132] In this embodiment, by converting the landing page of each product into a recommendation link, when the recommendation platform recommends a product to a user, the user's behavior data can be recorded through the recommendation link, the recommendation effect data of the recommendation platform can be tracked, and the diversity of user information data can be enriched.

[0133] In summary, the product information recommendation method provided in the embodiment of the present application can recall different product recall sets according to the recommendation data of different recommendation platforms, which can achieve a more accurate recommendation effect. At the same time, by transferring the product landing page, it can facilitate the use of the recommendation platform, and track and record the user's click data, etc., and then use a personalized sorting algorithm to recommend personalized products to users. Finally, products can be sorted according to different preset rules, which improves the flexibility of the entire sorting process and enhances the stability of product information recommendations.

[0134] The following is an embodiment of the device of the present application, which can be used to execute the embodiment of the method of the present application. For details not disclosed in the embodiment of the device of the present application, please refer to the embodiment of the method of the present application.

[0135] Figure 8 This is a structural diagram of a first embodiment of a device for pushing product information provided in an embodiment of the present application, such as Figure 8 As shown, the device can be integrated in the remote terminal, or it can be independent of the remote terminal and work in conjunction with the remote terminal to implement the technical solution of the present application, such as Figure 8 As shown, the push device 80 includes a data acquisition module 801, a recall module 802, a determination module 803 and a recommendation module 804.

[0136] The data acquisition module 801 is used to acquire recommendation data from the recommendation platform.

[0137] The recall module 802 is used to recall products according to the recommended data to obtain a recall set.

[0138] The determination module 803 is used to determine at least one product to be recommended from the recalled set according to a preset rule to obtain a recommended set.

[0139] The recommendation module 804 is used to send recommendation information to the recommendation platform according to the recommendation set.

[0140] Among them, the recommendation data is used to indicate the recommendation effects of different products. The recall set includes the identifications and product information of multiple products. The product information of each product includes value information at different value attributes, recommendation result indicators and at least one information of the acquisition method. The recommendation set includes the identification and product information of at least one product to be recommended. The recommendation information is used to instruct the recommendation platform to recommend at least one product to be recommended.

[0141] Optionally, in some embodiments, the recall module 802 can be specifically used to perform a scheduled asynchronous recall of products according to preset scheduled tasks and recommendation data to obtain a product candidate set, and filter the product candidate set according to preset business rules to obtain a recall set.

[0142] Among them, the preset business rules include at least one of recommendation fee rules and product discount rules.

[0143] Optionally, in some embodiments, the preset rule may be a default sequence. If the preset rule is a default sequence, the determination module 803 may be specifically used to store the recall set in a preset database, and obtain the default sequence of products in the recall set stored in the preset database, determine at least one product to be recommended according to the default sequence, and obtain a recommended set.

[0144] Optionally, in some embodiments, the preset rule may be a preset sorting rule. If the preset rule is a preset sorting rule, the determination module 803 may be specifically used to sort the products in the recall set according to the preset sorting rule to obtain a product arrangement order, and determine at least one product to be recommended according to the product arrangement order to obtain a recommendation set.

[0145] Optionally, in some embodiments, the preset sorting rules include a first sorting rule and a second sorting rule, the first sorting rule includes a preset sorting algorithm, and the second sorting rule includes at least one of a product price sorting rule, a product recommendation cost sorting rule, and a product evaluation information sorting rule. The determination module 803 can be specifically used to sort products according to the first sorting rule and the product information of the recall set to obtain a product arrangement order; or to sort products according to the second sorting rule and the product information of the recall set to obtain a product arrangement order.

[0146] Optionally, in some embodiments, the preset sorting algorithm includes a gradient boosting decision tree algorithm and a logistic regression algorithm. The determination module 803 can be specifically used to obtain user tags, which include user attribute tags and user behavior tags, and input the user tags and product information of the products in the recall set into the preset sorting algorithm to sort the products and obtain the product arrangement order.

[0147] Optionally, in some embodiments, the determination module 803 can be specifically used to obtain user attribute data and user behavior data through online logs, real-time messages and offline calculations, and encode the user attribute data to obtain user attribute labels, discretize the user behavior data to obtain user behavior labels, and store the user attribute labels and user behavior labels in a preset database, and read the preset database to obtain user labels.

[0148] Exemplarily, in any of the above embodiments, the recommendation device further includes a testing module.

[0149] The test module is used to configure the preset sorting algorithm according to the preset algorithm configuration, test the preset sorting algorithm after the algorithm configuration to obtain test data, and optimize the configuration of the preset sorting algorithm according to the test data.

[0150] Exemplarily, in any of the above embodiments, the recommendation device further includes a cache module.

[0151] The cache module is used to obtain identification parameters and cache the recommendation set to a preset database according to the identification parameters. The identification parameters include a scene identification, a user identification and test data.

[0152] Optionally, in some embodiments, the recommendation module 804 can be specifically used to obtain the landing page of each product in the recommendation set, and perform link conversion on the landing page of each product in the recommendation set to obtain a converted recommendation link, and obtain recommendation information based on the recommendation link and the product information of the product, and send the recommendation information to the recommendation platform.

[0153] The device provided in the embodiment of the present application can be used to perform Figures 2 to 7The implementation principle and technical effect of the method in the illustrated embodiment are similar and will not be described in detail here.

[0154] It should be noted that it should be understood that the division of the various modules of the above device is only a division of logical functions. In actual implementation, they can be fully or partially integrated into one physical entity, or physically separated. Moreover, these modules can be all implemented in the form of software calling through processing elements; they can also be all implemented in the form of hardware; some modules can also be implemented in the form of software calling through processing elements, and some modules can be implemented in the form of hardware.

[0155] As an example, Fig. 9 The structural diagram of the second embodiment of the device for recommending product information provided by the present application is as follows: Fig. 9As shown, the recommendation device can be divided into module 1, module 2 and module 3. The steps executed by module 1 include step S11, step S12 and subsequent steps S13 to S16, the steps executed by module 2 include steps S21 to S32, and the steps executed by module 3 include step S33. Among them, first module 1 is an external service, which is used to request input parameters, scene identification and intercept invalid requests. Then module 2 performs asynchronous recall, recommendation API and sorting functions. Specifically, when performing asynchronous recall, based on different recommendation platforms, product information is loaded regularly and products that do not meet business rules are filtered to obtain a recall set with a default sequence. Then the TOP N products in the recall set are stored in the Redis database. When performing the recommendation API, relevant input parameters are parsed. When performing the sorting function, it is determined whether custom sorting is required based on the operator's request. The custom sorting can be the second sorting rule mentioned above. If the custom sorting conditions are met, custom sorting is performed directly in the memory, and then the sorting results are cached with the scene identification, sorting field, and ascending and descending dimensions. If the custom sorting conditions are not met, sorting is performed according to the first sorting rule mentioned above. Before this, AB testing is required to configure the preset sorting algorithm, and then sorting is performed according to the user tag and the configured preset sorting algorithm. Finally, the sorting results are cached with the scene identification, AB parameters, and user unique identification parameters. Finally, the algorithm sorting results and the default sequence of the recall set are summarized, and the recommended set, i.e., SKU, is returned according to the paging information. ID list, as well as the granularity of the AB experimental parameters of the AB test, and then module 3 will transfer the URL link of the product's landing page to obtain the product's conversion link, and finally return to module 1. If module 2 of the entire recommendation device can work normally, the result returned to module 1 is normal. Module 1 supplements the product information according to the SKU ID list in the paging information, including whether the product has a coupon, etc., and finally obtains the recommendation information as the recommendation result and returns it to the recommendation platform. If module 2 cannot work normally, the above-mentioned sorting function may not be able to directly obtain the TOP N products stored in the redis database at this time, and supplement the product information of these products to obtain the recommendation information as the recommendation result and return it to the recommendation platform.

[0156] Optional, Fig.10 A schematic diagram of a recommended device for product information provided in an embodiment of the present application. The recommended device may be a courier cabinet or a mobile terminal, such as Fig.10 As shown, the recommendation device 100 includes a memory 1001 and at least one processor 1002;

[0157] The memory 1001 stores computer-executable instructions; the processor 1002 executes the computer-executable instructions stored in the memory 1001 , so that the processor 1002 performs the method steps as described above.

[0158] Exemplarily, the memory 1001 may be connected to the processor 1002 via a bus 1003 . The processor 1002 may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), and the like.

[0159] Optionally, an embodiment of the present application further provides a computer-readable storage medium having a computer program stored thereon, and the computer program implements the above-mentioned method steps when executed by a processor.

[0160] In the present application, "at least one" means one or more, and "plurality" means two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone, where A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship; in the formula, the character " / " indicates that the previous and next associated objects are in a "division" relationship. "At least one of the following items" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, c can be single or multiple.

[0161] It is to be understood that the various numerical numbers involved in the embodiments of the present application are only for the convenience of description and are not intended to limit the scope of the embodiments of the present application. In the embodiments of the present application, the size of the sequence number of each process does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0162] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for recommending product information, characterized in that: include: Acquire recommendation data from a recommendation platform, where the recommendation data is used to indicate recommendation effects of different products; Performing product recall according to the recommendation data to obtain a recall set, wherein the recall set includes identifications and product information of multiple products, and the product information of each product includes value information at different value attributes, a recommendation result indicator, and at least one of an acquisition method; According to a preset rule, at least one product to be recommended is determined from the recalled set to obtain a recommendation set, wherein the recommendation set includes an identifier and product information of the at least one product to be recommended, and the preset rule includes a default sequence or a preset sorting rule; Sending recommendation information to the recommendation platform according to the recommendation set, wherein the recommendation information is used to instruct the recommendation platform to recommend the at least one product to be recommended; The step of determining at least one product to be recommended from the recalled set according to a preset rule to obtain a recommended set includes: Storing the recalled set in a preset database; Acquire a default sequence of products in a recalled set stored in the preset database, determine at least one product to be recommended according to the default sequence, and obtain a recommended set; or, According to the preset sorting rule, the products in the recall set are sorted to obtain a product arrangement order; the preset sorting rule includes a first sorting rule and a second sorting rule, the first sorting rule includes a preset sorting algorithm, and the second sorting rule includes at least one of a product price sorting rule, a product recommendation fee sorting rule, and a product evaluation information sorting rule; At least one product to be recommended is determined according to the product arrangement order to obtain a recommendation set.

2. The method according to claim 1, characterized in that The step of recalling the product according to the recommendation data to obtain a recall set includes: Performing a timed asynchronous recall of products according to preset timed tasks and the recommendation data to obtain a set of product candidates; The product candidate set is filtered according to preset business rules to obtain the recalled set, wherein the preset business rules include at least one of a recommended fee rule and a product discount rule.

3. The method according to claim 1, characterized in that Sorting the products in the recall set according to the preset sorting rule to obtain a product arrangement order includes: Sort the products according to the first sorting rule and the product information of the recalled set to obtain a product arrangement order; or, The products are sorted according to the second sorting rule and the product information of the recalled set to obtain a product arrangement order.

4. The method according to claim 3, characterized in that The step of sorting the products according to the first sorting rule and the product information of the recalled set to obtain a product arrangement order includes: Obtaining user tags, wherein the user tags include user attribute tags and user behavior tags; The user tags and the product information of the products in the recall set are input into the preset sorting algorithm to sort the products to obtain a product arrangement order. The preset sorting algorithm includes a gradient boosting decision tree algorithm and a logistic regression algorithm.

5. The method according to claim 4, characterized in that The obtaining of the user tag comprises: Obtain user attribute data and user behavior data through online logs, real-time messages, and offline calculations; Encoding the user attribute data to obtain a user attribute tag; Discretizing the user behavior data to obtain a user behavior label; The user attribute tag and the user behavior tag are stored in a preset database, and the preset database is read to obtain the user tag.

6. The method according to claim 4, characterized in that Before inputting the user tag and the product information of the products in the recall set into the preset sorting algorithm to sort the products and obtain the product arrangement order, the method further includes: Performing algorithm configuration on the preset sorting algorithm according to the preset algorithm configuration; Testing the preset sorting algorithm after algorithm configuration to obtain test data; The preset sorting algorithm is configured and optimized according to the test data.

7. The method according to claim 6, characterized in that After determining at least one product to be recommended according to the product arrangement order and obtaining a recommendation set, the method further comprises: Obtain identification parameters, and cache the recommendation set to a preset database according to the identification parameters, wherein the identification parameters include a scene identification, a user identification, and test data.

8. The method according to claim 1, characterized in that The sending the recommendation information to the recommendation platform according to the recommendation set includes: Obtain the landing page of each product in the recommendation set; Performing link conversion on the landing page of each product in the recommendation set to obtain a converted recommendation link, wherein the recommendation link is used to collect recommendation data of the recommendation platform; The recommendation information is obtained according to the recommendation link and the product information of the product, and the recommendation information is sent to the recommendation platform.

9. A device for recommending product information, characterized in that: include: A data acquisition module, used to acquire recommendation data from the recommendation platform, wherein the recommendation data is used to indicate the recommendation effects of different products; A recall module, configured to recall products according to the recommendation data to obtain a recall set, wherein the recall set includes identifiers and product information of multiple products, and the product information of each product includes value information at different value attributes, a recommendation result indicator, and at least one of an acquisition method; A determination module, configured to determine at least one product to be recommended from the recalled set according to a preset rule to obtain a recommendation set, wherein the recommendation set includes an identifier and product information of the at least one product to be recommended, and the preset rule includes a default sequence or a preset sorting rule; A recommendation module, configured to send recommendation information to the recommendation platform according to the recommendation set, wherein the recommendation information is used to instruct the recommendation platform to recommend the at least one product to be recommended; The determination module is specifically configured to store the recall set in a preset database; obtain a default sequence of products in the recall set stored in the preset database, determine at least one product to be recommended according to the default sequence, and obtain a recommendation set; or, According to the preset sorting rules, the products in the recall set are sorted to obtain a product arrangement order; the preset sorting rules include a first sorting rule and a second sorting rule, the first sorting rule includes a preset sorting algorithm, and the second sorting rule includes at least one of a product price sorting rule, a product recommendation fee sorting rule and a product evaluation information sorting rule; according to the product arrangement order, at least one product to be recommended is determined to obtain a recommendation set.

10. A device for recommending product information, characterized in that: comprising a memory and at least one processor; The memory stores computer-executable instructions; The at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor performs the method according to any one of claims 1 to 8.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that: The computer-readable storage medium stores computer instructions, which are used to implement the method according to any one of claims 1 to 8 when executed by a processor.

Images