Recommendation method, device, computer equipment and storage medium based on user portrait

Abstract

The present invention provides a recommendation method, device, computer equipment and storage medium based on user portrait, the method comprising obtaining user behavior data; parsing the user behavior data to obtain a user feature vector; constructing a user portrait based on the user feature vector; extracting the first several products arranged according to a preset rule from the user portrait to obtain several pre-selected products; obtaining product classification data; associating the product classification data with each of the pre-selected products to obtain the product classification corresponding to each of the pre-selected products in the product classification data; and pushing product messages according to the product classification. By constructing a user portrait and associating the user portrait with the product classification data, the product classification that the user is interested in is determined, and the product message corresponding to the product classification is pushed to the user, so that the product message can be pushed accurately, and the experience is better.

Description

Technical Field

[0001] The present invention relates to the technical field of commodity push, and in particular to a recommendation method, device, computer equipment and storage medium based on user portrait. Background Art

[0002] With the continuous development of the Internet, especially the rapid development of mobile Internet, various applications are being developed to provide better user experience and targeted product push. Applications can push messages based on user subscriptions or preferences.

[0003] Push messages are usually subscribed manually by users, and cannot be intelligently recommended, which results in users only paying attention to content they are already interested in, and not being able to push messages that they may be interested in. For some products that are simply obtained based on the user's interests, the push is inaccurate due to the relatively simple calculation process of the products of interest, resulting in poor user perception. Summary of the invention

[0004] Based on this, it is necessary to provide a recommendation method, device, computer equipment and storage medium based on user portrait to address the above technical problems.

[0005] A recommendation method based on user portrait, comprising:

[0006] Obtain user behavior data;

[0007] Analyzing the user behavior data to obtain a user feature vector;

[0008] Building a user portrait based on the user feature vector;

[0009] Extracting the first several products arranged according to a preset rule from the user portrait to obtain several pre-selected products;

[0010] Get product classification data;

[0011] Associating the product classification data with each of the pre-selected products, and obtaining the product classification corresponding to each of the pre-selected products in the product classification data;

[0012] Push product news according to the product classification.

[0013] In one embodiment, the step of constructing a user portrait based on the user feature vector includes:

[0014] Based on the user feature vector, the user portrait is constructed using the TF-IDF algorithm.

[0015] In one embodiment, the step of constructing the user portrait based on the user feature vector using the TF-IDF algorithm includes:

[0016] Based on the user feature vector, obtaining weights of various products purchased by the user;

[0017] Based on the weights of various products purchased by the user, the user portrait is constructed using the TF-IDF algorithm.

[0018] In one embodiment, the step of extracting the first several products arranged according to a preset rule from the user portrait to obtain several pre-selected products includes:

[0019] Based on the weights of various products in the user portrait, the first several products are extracted from the user portrait in descending order of weight, and several pre-selected products are obtained.

[0020] In one embodiment, the step of pushing product information according to the product classification includes:

[0021] Generate a product push list according to the product classification obtained from the product classification data;

[0022] Push the product push list.

[0023] In one embodiment, the user behavior data obtained includes user registration data, user comment data, user click data and user geographic location data.

[0024] A recommendation device based on user portrait, comprising:

[0025] User behavior acquisition module, used to acquire user behavior data;

[0026] A user feature vector acquisition module, used to analyze the user behavior data and obtain a user feature vector;

[0027] A user portrait building module, used to build a user portrait based on the user feature vector;

[0028] A pre-selected product extraction module is used to extract the first several products arranged according to a preset rule from the user portrait to obtain several pre-selected products;

[0029] Product classification data acquisition module, used to obtain product classification data;

[0030] A product classification acquisition module, used to associate the product classification data with each of the pre-selected products, and obtain the product classification corresponding to each of the pre-selected products in the product classification data;

[0031] The product message push module is used to push product messages according to the product classification.

[0032] In one of the embodiments, the user portrait construction module is further used to construct the user portrait based on the user feature vector using a TF-IDF algorithm.

[0033] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and wherein the processor implements the following steps when executing the computer program:

[0034] Obtain user behavior data;

[0035] Analyzing the user behavior data to obtain a user feature vector;

[0036] Building a user portrait based on the user feature vector;

[0037] Extracting the first several products arranged according to a preset rule from the user portrait to obtain several pre-selected products;

[0038] Get product classification data;

[0039] Associating the product classification data with each of the pre-selected products, and obtaining the product classification corresponding to each of the pre-selected products in the product classification data;

[0040] Push product news according to the product classification.

[0041] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the following steps:

[0042] Obtain user behavior data;

[0043] Analyzing the user behavior data to obtain a user feature vector;

[0044] Building a user portrait based on the user feature vector;

[0045] Extracting the first several products arranged according to a preset rule from the user portrait to obtain several pre-selected products;

[0046] Get product classification data;

[0047] Associating the product classification data with each of the pre-selected products, and obtaining the product classification corresponding to each of the pre-selected products in the product classification data;

[0048] Push product news according to the product classification.

[0049] The above-mentioned user portrait-based recommendation method, device, computer equipment and storage medium construct a user portrait and associate the user portrait with product classification data, thereby determining the product category that the user is interested in, and pushing the product message corresponding to the product category to the user, so that the product message can be pushed accurately, providing a better experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 A schematic diagram of an application scenario of a recommendation method based on user portrait in an embodiment;

[0051] Figure 2 is a flowchart of a recommendation method based on user portrait in one embodiment;

[0052] Figure 3 is a structural block diagram of a recommendation device based on user portrait in one embodiment;

[0053] Figure 4 is an internal structure diagram of a computer device in one embodiment;

[0054] Figure 5 The figure is a schematic diagram of the implementation process of a recommendation method based on user portrait in one embodiment. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0056] Embodiment 1

[0057] The recommendation method based on user portrait provided in this application can be applied to Figure 1In the application environment shown. Among them, the computer 102 communicates with the server 104 through the network. Among them, the terminal 102 can be but not limited to various personal computers, servers, laptops, smart phones, tablet computers and portable wearable devices, and the server 104 can be implemented by an independent server or a server cluster composed of multiple servers. The application software runs on the terminal 102, and the user accesses the server 104 through the application software on the terminal 102, registers the user account, and browses, collects, comments and purchases the goods displayed on the server 104. The server 104 obtains the user behavior data such as registration, browsing, collection, comment and purchase of the user, analyzes the user behavior data, and obtains the user feature vector; based on the user feature vector, constructs a user portrait; extracts the first several products arranged according to the preset rules from the user portrait, and obtains several pre-selected products; obtains product classification data; associates the product classification data with each of the pre-selected products, and obtains the product classification corresponding to each of the pre-selected products in the product classification data; pushes product messages to the terminal 102 according to the product classification. When users browse goods, they can be accurately pushed product messages, making the experience better.

[0058] Embodiment 2

[0059] In this embodiment, Figure 2 As shown, a recommendation method based on user portrait is provided, which includes:

[0060] Step 210, obtaining user behavior data.

[0061] In this step, the user behavior data is collected. The user behavior includes: the user's registration behavior, the operation behavior of the products in the application, and the operation behavior of the products in the application includes clicking, browsing, commenting, collecting, purchasing, etc. The comment can be entered as text information or clicked on the star rating or rating to comment. The user behavior data is the statistical data of these behaviors or the input data. For example, the rating of the evaluation, the number of purchases, the purchase price, the browsing time, the number of clicks, etc.

[0062] In one embodiment, the user behavior data obtained includes user registration data, user comment data, user click data and user geographic location data.

[0063] In this embodiment, user registration data refers to user registration information, including user's personal information, such as user's age, gender, location, occupation, interests, education, etc. User review data includes user evaluation data, such as users' good, bad and medium reviews of products. User click data includes the frequency and number of clicks on products. User geographic location data refers to user geographic location information. It should be understood that the user behavior data is the data source for building user portraits. In this application, user data is collected and product data is updated through user behavior. The collected user data includes user registration information, obtaining user comments, obtaining user click information through embedding points, user geographic location, etc.; updating product data: obtaining the number of user clicks on a product to update the popularity of the product.

[0064] In one embodiment, after acquiring the user behavior data, the process further includes storing the user behavior data.

[0065] In this embodiment, the collected data is stored in a disk or memory. Depending on the size of the user behavior data, it can be stored in a stand-alone or distributed file system. Data is divided into structured data and unstructured data. Structured data mainly exists in relational data models, such as MySQL and Hive, and unstructured data mainly exists in non-relational data models, such as Redis and Hbase.

[0066] Step 220: parse the user behavior data to obtain a user feature vector.

[0067] Specifically, each user can be represented by a vector, which contains various behavior data of the user. By converting the user behavior data into a user feature vector, the user feature vector reflects the user's interest in different products. In this embodiment, the user behavior data is parsed to obtain the weight value of the user's interest in the product, and based on the weight value, a user feature vector containing the weight value is obtained.

[0068] In one embodiment, the TF-IDF algorithm is used to parse the user behavior data to obtain a structured representation of user features, namely, a user feature vector.

[0069] Specifically, TF-IDF (term frequency–inverse document frequency) is a common weighting technique used in information retrieval and text mining. TF-IDF is a statistical method used to assess the importance of a word to a set of documents or one of the documents in a corpus. The importance of a word increases in direct proportion to the number of times it appears in a document, but it decreases in inverse proportion to the frequency of its occurrence in the corpus. Various forms of TF-IDF weighting are often used by search engines as a measure or rating of the relevance between documents and user queries. In addition to TF-IDF, search engines on the Internet also use rating methods based on link analysis to determine the order in which documents appear in search results.

[0070] Step 230: construct a user portrait based on the user feature vector.

[0071] In one embodiment, the step of constructing a user portrait based on the user feature vector includes: constructing the user portrait based on the user feature vector using a TF-IDF algorithm.

[0072] In one embodiment, the step of constructing the user portrait based on the user feature vector using the TF-IDF algorithm includes: obtaining the weights of various products purchased by the user based on the user feature vector; and constructing the user portrait using the TF-IDF algorithm based on the weights of various products purchased by the user. In this embodiment, the weights of the products purchased by the user are used to reflect the user's interest in the products. In this way, a user portrait reflecting the user's interest in different products can be constructed through the product weights.

[0073] The build process is:

[0074] In this embodiment, based on user behavior data, the TF-IDF algorithm is used to obtain a structured representation of user characteristics, based on which a user portrait is constructed. Assuming that the user has purchased some products, the detailed process of constructing a user portrait is as follows:

[0075] (1) Let the set of all users be D = {d1, d2, ..., d m}, the set of all products purchased by these users is T = {t1, t2, ..., t n}, that is, m users purchased n different products.

[0076] (2) Represent each user as a vector: d j =(w 1j ,w 2j ,...,w nj ), where w 1jrepresents the weight of the first product for user j. The larger the weight, the more interested the user is in the product. Find the weight w of k products for user j k j The formula is:

[0077]

[0078]

[0079] Among them, TF(t k ,d j ) indicates user d j Buy Products k The number of times, n k represents the total number of times all users purchase the kth product; Formula (2) means to normalize the algorithm result and return the representation vectors of different users to the same level.

[0080] Step 240, extracting the first several products arranged according to a preset rule from the user portrait to obtain several pre-selected products.

[0081] In this step, the products are arranged according to preset rules, which can reflect the user's attention and interest in the products. The products in the user portrait are arranged according to preset rules, and several products at the top of the arrangement are extracted as pre-selected products.

[0082] Specifically, the user profile of each user includes the user's interest in different products, which can be reflected by user behavior data such as the number of purchases, the number of views, the duration of views, and the time of collection. Therefore, several products with the highest interest levels are extracted from the user profile as pre-selected products, so that the extracted products are the products that the user is most interested in.

[0083] In one embodiment, the step of extracting the first several products arranged according to preset rules from the user portrait to obtain several pre-selected products includes: based on the weights of each type of product in the user portrait, extracting the first several products from the user portrait in descending order of weight to obtain several of the pre-selected products.

[0084] In this embodiment, since the user portrait is constructed based on the weight of the product to the user, the weight of each product to the user can be obtained through the constructed user portrait, and the weight value of the product can be obtained. Then, the products ranked in the top of the ranking are extracted as pre-selected products according to the descending order of the weight value. In this way, the extracted products are the products that the user is most interested in.

[0085] Step 250, obtaining product classification data.

[0086] Specifically, the product classification data can also be called a product classification table. The product classification data records the classification of various products. For example, a certain mobile phone has a subcategory of mobile phones or communication tools and a major category of digital products. The product classification data is a pre-established data table.

[0087] Step 260: Associating the product classification data with each of the pre-selected products, and obtaining the product classification corresponding to each of the pre-selected products in the product classification data.

[0088] In this embodiment, the product classification data is associated with the user portrait, so that the product classification corresponding to the category of the pre-selected product can be obtained from the product classification data. For example, if the extracted pre-selected product is a certain model of mobile phone, the product classification corresponding to the category of the pre-selected product obtained from the product classification data is a mobile phone or a digital product.

[0089] Step 270: Push product information according to the product classification.

[0090] In this step, after determining the product category in the product classification data according to the pre-selected products, the product message corresponding to the product category is pushed to the user's client, so that the product can be pushed to the user accurately, improving the user's usage perception and making the user experience better.

[0091] In one embodiment, the step of pushing product messages according to the product classification includes: generating a product push list according to the product classification obtained from the product classification data; and pushing the product push list.

[0092] In this embodiment, since the number of product categories obtained from the product classification data can be multiple, in order to facilitate the push of products corresponding to these product categories, a product push list containing each product category is generated, and the product push list is pushed to the user, so that multiple product messages can be pushed at the same time, thereby improving the push efficiency.

[0093] Embodiment 3

[0094] In this embodiment, see Figure 5 , Figure 5 The "product" in the description is the specific object to be pushed, and the "push message" in this application can also be specifically understood as "push product". For the sake of convenience, the following description uniformly refers to "message" as "product".

[0095] Users can perform some actions on products, such as following, purchasing, and collecting, etc. These actions are collectively referred to as "purchases" below.

[0096] 1. Data collection: The data source for building user portraits. The intelligent push system collects user data and updates product data through user behavior. Collect user data: user registration information, obtain user comments, obtain user click information and user geographic location through embedding points; update product data: obtain the number of user clicks on a product to update the popularity of the product.

[0097] 2. Data storage: The collected data is stored in disk or memory. Depending on the size of the data, it can be stored in a single machine or distributed file system. Data is divided into structured data and unstructured data. Structured data mainly exists in relational data models (MySQL, Hive), and unstructured data mainly exists in non-relational data models (Redis, Hbase).

[0098] 3. User portrait construction: Based on the user data stored in the database, the TF-IDF algorithm is used to obtain a structured representation of user features (i.e., user portrait). Assuming that the user has purchased some products, the detailed process of constructing a user portrait is as follows:

[0099] (1) Let the set of all users be D = {d1, d2, ..., d m}, the set of all products purchased by these users is T = {t1, t2, ..., t n}, that is, m users purchased n different products.

[0100] (2) Represent each user as a vector: d j =(w 1j ,w 2j ,...,w nj ), where w 1j represents the weight of the first product for user j. The larger the weight value, the more interested the user is in the product. Find the weight w of k products for user j k j The formula is:

[0101]

[0102]

[0103] Among them, TF(t k ,d j ) indicates user d j Buy Products k The number of times, n k represents the total number of times all users purchase the kth product; Formula (2) means to normalize the algorithm result and return the representation vectors of different users to the same level.

[0104] 4. Product classification: Classify all products, such as classifying Apple into food-fruit category, Xiaomi 10 into digital product-mobile phone category, etc., and group these data together to form a structured relational data table.

[0105] 5. Association: Take out the top-ranked products in the user portrait and associate them with the data table in step 4 to obtain other products of the same category that should be recommended to the user.

[0106] 6. Recommendation: Sort the results in step 5, generate a recommendation list for the user, and push the recommendation list to the user.

[0107] Embodiment 4

[0108] In this embodiment, Figure 3 As shown, a recommendation device based on user portrait is provided, including:

[0109] User behavior acquisition module 310, used to acquire user behavior data;

[0110] A user feature vector acquisition module 320, used to parse the user behavior data and obtain a user feature vector;

[0111] A user portrait construction module 330, used to construct a user portrait based on the user feature vector;

[0112] A pre-selected product extraction module 340 is used to extract the first several products arranged according to a preset rule from the user portrait to obtain several pre-selected products;

[0113] A product classification data acquisition module 350, for acquiring product classification data;

[0114] A product classification acquisition module 360, configured to associate the product classification data with each of the pre-selected products, and acquire the product classification corresponding to each of the pre-selected products in the product classification data;

[0115] The product message push module 370 is used to push product messages according to the product classification.

[0116] In one embodiment, the user portrait construction module is further used to construct the user portrait based on the user feature vector using the TF-IDF algorithm.

[0117] In one embodiment, the user portrait building module includes:

[0118] A product weight acquisition unit, used to acquire the weights of various products purchased by the user based on the user feature vector;

[0119] The user portrait construction unit is used to construct the user portrait based on the weights of various products purchased by the user using the TF-IDF algorithm.

[0120] In one embodiment, the pre-selected product extraction module is further used to extract the first several products from the user portrait based on the weights of various types of products in the user portrait, in descending order of weight, to obtain several of the pre-selected products.

[0121] In one embodiment, the product message push module includes:

[0122] A product push list generation loop, used to generate a product push list according to the product classification obtained from the product classification data;

[0123] The product push list pushing unit is used to push the product push list.

[0124] In one embodiment, the user behavior data obtained includes user registration data, user comment data, user click data and user geographic location data.

[0125] For the specific definition of the recommendation device based on user portrait, please refer to the definition of the recommendation method based on user portrait above, which will not be repeated here. Each unit in the above-mentioned recommendation device based on user portrait can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned units can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above-mentioned units.

[0126] Embodiment 5

[0127] In this embodiment, a computer device is provided. Its internal structure diagram can be shown as follows: Figure 4As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program, and the non-volatile storage medium is deployed with a database, which is used to store user behavior data and user portraits. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used to communicate with other computer devices that deploy application software. When the computer program is executed by the processor, a recommendation method based on user portraits is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a button, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0128] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0129] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0130] Obtain user behavior data;

[0131] Analyzing the user behavior data to obtain a user feature vector;

[0132] Building a user portrait based on the user feature vector;

[0133] Extracting the first several products arranged according to a preset rule from the user portrait to obtain several pre-selected products;

[0134] Get product classification data;

[0135] Associating the product classification data with each of the pre-selected products, and obtaining the product classification corresponding to each of the pre-selected products in the product classification data;

[0136] Push product news according to the product classification.

[0137] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0138] Based on the user feature vector, the user portrait is constructed using the TF-IDF algorithm.

[0139] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0140] Based on the user feature vector, obtaining weights of various products purchased by the user;

[0141] Based on the weights of various products purchased by the user, the user portrait is constructed using the TF-IDF algorithm.

[0142] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0143] Based on the weights of various products in the user portrait, the first several products are extracted from the user portrait in descending order of weight, and several pre-selected products are obtained.

[0144] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0145] Generate a product push list according to the product classification obtained from the product classification data;

[0146] Push the product push list.

[0147] Embodiment 6

[0148] In this embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0149] Obtain user behavior data;

[0150] Analyzing the user behavior data to obtain a user feature vector;

[0151] Building a user portrait based on the user feature vector;

[0152] Extracting the first several products arranged according to a preset rule from the user portrait to obtain several pre-selected products;

[0153] Get product classification data;

[0154] Associating the product classification data with each of the pre-selected products, and obtaining the product classification corresponding to each of the pre-selected products in the product classification data;

[0155] Push product news according to the product classification.

[0156] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0157] Based on the user feature vector, the user portrait is constructed using the TF-IDF algorithm.

[0158] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0159] Based on the user feature vector, obtaining weights of various products purchased by the user;

[0160] Based on the weights of various products purchased by the user, the user portrait is constructed using the TF-IDF algorithm.

[0161] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0162] Based on the weights of various products in the user portrait, the first several products are extracted from the user portrait in descending order of weight, and several pre-selected products are obtained.

[0163] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0164] Generate a product push list according to the product classification obtained from the product classification data;

[0165] Push the product push list.

[0166] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0167] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0168] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A recommendation method based on user portrait, characterized in that: include: Acquire user behavior data, wherein the user behavior data includes user click data on products; Analyzing the user behavior data to obtain a user feature vector; Constructing a user portrait based on the user feature vector includes: constructing the user portrait based on the user feature vector using a TF-IDF algorithm, the steps including: Based on the user feature vector, obtaining weights of various products purchased by the user; Based on the weights of the various products purchased by the user, the TF-IDF algorithm is used to construct the user portrait, using the following expression: Among them, TF(t k ,d j ) for user d j Buy Products k The number of times; n k The total number of times all users purchased the kth product; w k j is the weight of k products for user j; D = {d1, d2, ..., d m } is the set of all users; T = {t1, t2, ..., t n } is the collection of all products purchased by the user; d j =(w 1j ,w 2j ,...,w nj ) is represented by a vector for each user; w 1j is the weight of the first product for user j; Extracting the first several products arranged according to a preset rule from the user portrait to obtain several pre-selected products, including: extracting the first several products from the user portrait in descending order of weight based on the weights of various types of products in the user portrait to obtain several pre-selected products; Get product classification data; Associating the product classification data with each of the pre-selected products, and obtaining the product classification corresponding to each of the pre-selected products in the product classification data; Push product news according to the product classification.

2. The method according to claim 1, characterized in that The step of pushing product information according to the product classification includes: Generate a product push list according to the product classification obtained from the product classification data; Push the product push list.

3. The method according to claim 1, characterized in that The user behavior data obtained also includes user registration data, user comment data and user geographic location data.

4. A recommendation device based on user portrait, characterized in that: include: A user behavior acquisition module, used to acquire user behavior data, wherein the user behavior data includes user click data on products; A user feature vector acquisition module, used to analyze the user behavior data and obtain a user feature vector; The user portrait construction module is used to construct a user portrait based on the user feature vector, including: based on the user feature vector, using the TF-IDF algorithm to construct the user portrait, the steps include: Based on the user feature vector, obtaining weights of various products purchased by the user; Based on the weights of the various products purchased by the user, the TF-IDF algorithm is used to construct the user portrait, using the following expression: Among them, TF(t k ,d j ) for user d j Buy Products k The number of times; n k The total number of times all users purchased the kth product; w k j is the weight of k products for user j; D = {d1, d2, ..., d m } is the set of all users; T = {t1, t2, ..., t n } is the collection of all products purchased by the user; d j =(w 1j ,w 2j ,...,w nj ) is represented by a vector for each user; w 1j is the weight of the first product for user j; A pre-selected product extraction module is used to extract the first several products arranged according to a preset rule from the user portrait to obtain several pre-selected products, including: based on the weights of various products in the user portrait, extracting the first several products from the user portrait in descending order of weight to obtain several pre-selected products; Product classification data acquisition module, used to obtain product classification data; A product classification acquisition module, used to associate the product classification data with each of the pre-selected products, and obtain the product classification corresponding to each of the pre-selected products in the product classification data; The product message push module is used to push product messages according to the product classification.

5. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the user portrait-based recommendation method described in any one of claims 1 to 3 are implemented.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the user portrait-based recommendation method described in any one of claims 1 to 3 are implemented.

Images