Product recommendation method and device, equipment and storage medium

Abstract

The invention provides a product recommendation method and device, equipment and a storage medium, and can be applied to the field of artificial intelligence and the field of financial science and technology. The method comprises the following steps: in response to a received product recommendation request for a target object, determining multiple pieces of object information of the target object from an object information library; feature extraction and feature screening are conducted on the multiple pieces of object information, object feature sets corresponding to the multiple candidate products are obtained, and the object feature sets comprise multiple object features; inputting the object feature sets respectively corresponding to the plurality of candidate products into respective corresponding prediction models to obtain respective score values of the target object to the plurality of candidate products; determining similar objects of the target object from the plurality of preset objects based on the respective score values of the target object to the plurality of candidate products and the respective score values of the plurality of preset objects to the plurality of candidate products; and determining a to-be-recommended product for the target object based on the score values of the similar objects for the plurality of candidate products.

Description

Technical Field

[0001] The present disclosure relates to the fields of artificial intelligence and financial technology, and in particular to a product recommendation method, device, equipment and storage medium. Background Art

[0002] With the continuous development of Internet technology, based on the convenience of online shopping, more and more users choose online shopping. In order to provide users with a better service experience, companies usually determine the products that users may be interested in based on their product browsing or historical purchases, and recommend similar products to users so that users can purchase their favorite products.

[0003] In the process of realizing the concept of the present disclosure, the inventors discovered that there are at least the following problems in the related art: the product recommendation method in the related art may have the problem that the recommended products cannot meet the needs of users. Summary of the invention

[0004] In view of the above problems, the present disclosure provides product recommendations, devices, equipment, media and program products.

[0005] According to one aspect of the present disclosure, a product recommendation method is provided, comprising:

[0006] In response to receiving a product recommendation request for a target object, determining a plurality of object information of the target object from an object information library;

[0007] Extracting and screening features of multiple object information to obtain object feature sets corresponding to multiple candidate products, wherein the object feature set includes multiple object features;

[0008] Inputting the object feature sets corresponding to the multiple candidate products into the corresponding prediction models respectively, and obtaining the score values ​​of the target object for the multiple candidate products respectively;

[0009] Determine similar objects to the target object from the plurality of preset objects based on the respective scoring values ​​of the target object for the plurality of candidate products and the respective scoring values ​​of the plurality of preset objects for the plurality of candidate products;

[0010] Based on the ratings of multiple candidate products by similar objects, a product to be recommended for the target object is determined.

[0011] According to an embodiment of the present disclosure, the object feature sets corresponding to the multiple candidate products are respectively input into the corresponding prediction models to obtain the score values ​​of the target object for the multiple candidate products, including:

[0012] Inputting the object feature sets corresponding to the multiple candidate products into the corresponding prediction models, and obtaining the probability value of the target object holding each candidate product;

[0013] The probability value of the target object holding each candidate product is normalized to obtain the target object's respective score values ​​for multiple candidate products.

[0014] According to an embodiment of the present disclosure, based on the score values ​​of the target object for each of the multiple candidate products and the score values ​​of the multiple preset objects for each of the multiple candidate products, determining similar objects of the target object from the multiple preset objects includes:

[0015] Calculate the similarity between the score of the target object for each of the multiple candidate products and the score of each of the preset objects for each of the multiple candidate products to obtain the similarity between the target object and each of the preset objects;

[0016] Based on the similarity, a similar object to the target object is determined from a plurality of preset objects.

[0017] According to an embodiment of the present disclosure, similarity calculation is performed between the score value of the target object for each of the multiple candidate products and the score value of each of the preset objects for each of the multiple candidate products to obtain the similarity between the target object and each of the preset objects, including:

[0018] Determine a first target score value greater than a score threshold from the score values ​​of the target object for each of the plurality of candidate products;

[0019] Determining a second target evaluation value of the target candidate product by each preset object based on the product identifier of the target candidate product corresponding to the first target evaluation value;

[0020] Based on the product identification, a similarity calculation is performed between the first target rating value and the second target rating value of each preset object for the target candidate product to obtain the similarity between the target object and each preset object.

[0021] According to an embodiment of the present disclosure, determining a product to be recommended for a target object based on the score values ​​of a plurality of candidate products respectively given by similar objects in a similar object set includes:

[0022] The score of each candidate product by similar objects is preset to obtain the recommendation degree of each candidate product;

[0023] Sort the recommendation degree of each candidate product to obtain a ranking result;

[0024] Based on the ranking results, candidate products corresponding to a recommendation degree less than or equal to a preset order threshold are used as products to be recommended for the target object.

[0025] According to an embodiment of the present disclosure, feature extraction and feature screening are performed on multiple object information to obtain object feature sets corresponding to multiple candidate products, including:

[0026] Based on the information identifiers of the plurality of object information, the plurality of object information is divided into a plurality of target object information corresponding to the plurality of candidate products respectively;

[0027] Extracting features of multiple target object information corresponding to each of the multiple candidate products to obtain multiple initial feature sets;

[0028] Feature screening is performed on multiple initial feature sets respectively to obtain object feature sets corresponding to multiple candidate products respectively.

[0029] According to an embodiment of the present disclosure, the initial feature set includes a plurality of initial features; the plurality of initial feature sets are respectively subjected to feature screening to obtain object feature sets corresponding to a plurality of candidate products, including:

[0030] For the initial feature corresponding to each candidate product, based on the attribute information of the initial feature, determining a plurality of first intermediate features and a plurality of second intermediate features;

[0031] Performing variance calculation on the plurality of first intermediate features to obtain variance values ​​of the plurality of first intermediate features;

[0032] Filter out the first intermediate features whose variance values ​​are less than or equal to the variance threshold to obtain a first target feature set;

[0033] Calculating the correlation between the plurality of second intermediate features and the candidate products to obtain the correlation between the plurality of second target initial features;

[0034] Filter out the second intermediate features whose correlation degree is less than or equal to the correlation degree threshold to obtain a second target feature set;

[0035] Based on the first target feature set and the second target feature set corresponding to each of the plurality of candidate products, object feature sets corresponding to each of the plurality of candidate products are obtained.

[0036] Another aspect of the present disclosure provides a product recommendation device, comprising:

[0037] an object information determination module, configured to determine a plurality of object information of the target object from an object information library in response to receiving a product recommendation request for the target object;

[0038] An object feature set determination module is used to extract and screen features of multiple object information to obtain object feature sets corresponding to multiple candidate products, wherein the object feature set includes multiple object features;

[0039] A scoring value determination module is used to input the object feature sets corresponding to the multiple candidate products into the corresponding prediction models to obtain the target object's scoring values ​​for the multiple candidate products;

[0040] A similar object determination module, used to determine similar objects of a target object from a plurality of preset objects based on the respective scoring values ​​of the target object to the plurality of candidate products and the respective scoring values ​​of the plurality of preset objects to the plurality of candidate products;

[0041] The module for determining products to be recommended is used to determine products to be recommended for a target object based on the rating values ​​of multiple candidate products given by similar objects.

[0042] A third aspect of the present disclosure provides an electronic device, comprising: one or more processors; and a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors execute the above method.

[0043] The fourth aspect of the present disclosure also provides a computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, causes the processor to execute the above method.

[0044] The fifth aspect of the present disclosure also provides a computer program product, including a computer program, which implements the above method when executed by a processor.

[0045] According to the product recommendation method provided by the present disclosure, multiple object information of the target object is determined from the object information library, and feature extraction and feature screening are performed on the multiple object information respectively, so as to obtain the object feature set corresponding to each candidate product. The object feature set is input into the corresponding prediction model to obtain the score value of the target object for each of the multiple candidate products, and the similarity is calculated between the score value of the target object for each of the multiple candidate products and the score value of multiple preset objects for each of the multiple candidate products, so as to determine the similar objects of the target object, and starting from the score value of the similar objects for each candidate product, determine the product to be recommended corresponding to the target object. Since each candidate product has a corresponding prediction model, the object feature sets corresponding to each candidate product are input into the corresponding prediction models to more accurately determine the target object's respective ratings for multiple candidate products, that is, the target object's preference for multiple candidate objects. Based on the respective ratings of the multiple candidate products and the similarity of the respective ratings of the multiple preset objects, similar objects are determined for the target object, and the products to be recommended are determined based on the preferences of the similar objects for each candidate product. Therefore, the problem that the recommended products in the related technology cannot meet the user's needs is at least partially solved, and the technical effect of more accurately recommending products to users and improving the user's shopping experience is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The above contents and other objects, features and advantages of the present disclosure will become more apparent through the following description of the embodiments of the present disclosure with reference to the accompanying drawings, in which:

[0047] Figure 1 The following schematically shows an application scenario diagram of the product recommendation method, device, equipment and storage medium according to an embodiment of the present disclosure;

[0048] Figure 2 A flowchart of a product recommendation method according to an embodiment of the present disclosure is schematically shown;

[0049] Figure 3 A flowchart for determining a scoring value according to an embodiment of the present disclosure is schematically shown;

[0050] Figure 4 A schematic diagram of determining a score value based on a prediction model according to an embodiment of the present disclosure is schematically shown;

[0051] Figure 5 A schematic diagram schematically shows a product recommendation method according to an embodiment of the present disclosure;

[0052] Figure 6 A flowchart of a product recommendation method according to another embodiment of the present disclosure is schematically shown;

[0053] Figure 7 A structural block diagram of a product recommendation device according to an embodiment of the present disclosure is schematically shown; and

[0054] Figure 8 A block diagram of an electronic device suitable for implementing a product recommendation method according to an embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0055] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present disclosure. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure.

[0056] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. The terms "comprise", "include", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0057] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification, and should not be interpreted in an idealized or overly rigid manner.

[0058] When using expressions such as "at least one of A, B, and C, etc.", they should generally be interpreted according to the meaning of the expression commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0059] It should be noted that the product recommendation method and device disclosed herein can be used in the financial field to recommend financial products such as funds, and can also be used in any field other than the financial field, such as the field of computer technology. The application field of the product recommendation method and device disclosed herein is not limited.

[0060] In the technical solution of the present invention, the user information (including but not limited to user personal information, user image information, user device information, such as location information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved are all information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of relevant data comply with the relevant laws, regulations and standards of relevant countries and regions, take necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0061] During the research process, it was found that in related technologies, products are usually recommended to users based on their product browsing history or historical purchase history. However, the products recommended in this way have the problem of being too wide in range and not precise enough. Therefore, based on the above product recommendation method, it may not be possible to recommend products that meet user needs.

[0062] In the research process, we also found that the fixed customer base of the enterprise may have certain purchasing needs. Therefore, recommending products to this customer base can also achieve the effect of increasing the company's benefits. For example, for the bank's payroll customer base. Analysis shows that most of the payroll of this customer base flows to outside the bank, and there are few cases of purchasing products within the bank. There may also be customers with asset management needs among the payroll customer base, but because there is no purchase direction, it eventually leads to customer loss.

[0063] In view of this, an embodiment of the present disclosure provides a product recommendation method, comprising: in response to receiving a product recommendation request for a target object, determining multiple object information of the target object from an object information library; performing feature extraction and feature screening on the multiple object information to obtain object feature sets corresponding to multiple candidate products, wherein the object feature sets include multiple object features; inputting the object feature sets corresponding to the multiple candidate products into their respective corresponding prediction models to obtain respective scoring values ​​of the target object for the multiple candidate products; based on the respective scoring values ​​of the target object for the multiple candidate products and the respective scoring values ​​of multiple preset objects for the multiple candidate products, determining similar objects of the target object from the multiple preset objects; based on the scoring values ​​of the similar objects for the multiple candidate products, determining products to be recommended for the target object.

[0064] Figure 1 The application scenario diagram of the product recommendation method, device, equipment and storage medium according to the embodiments of the present disclosure is schematically shown.

[0065] like Figure 1 As shown, the application scenario 100 according to this embodiment may include terminal devices 101, 102, 103, a network 104, and a server 105. The network 104 is used to provide a medium for a communication link between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0066] Users can use terminal devices 101, 102, 103 to interact with server 105 through network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social platform software, etc. (only examples).

[0067] The terminal devices 101 , 102 , and 103 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, and desktop computers.

[0068] The server 105 may be a server that provides various services, such as a background management server (only an example) that provides support for websites browsed by users using the terminal devices 101, 102, and 103. The background management server may analyze and process the received data such as user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal device.

[0069] It should be noted that the product recommendation method provided in the embodiment of the present disclosure can generally be executed by the server 105. Accordingly, the product recommendation device provided in the embodiment of the present disclosure can generally be set in the server 105. The product recommendation method provided in the embodiment of the present disclosure can also be executed by a server or server cluster that is different from the server 105 and can communicate with the terminal devices 101, 102, 103 and / or the server 105. Accordingly, the product recommendation device provided in the embodiment of the present disclosure can also be set in a server or server cluster that is different from the server 105 and can communicate with the terminal devices 101, 102, 103 and / or the server 105.

[0070] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0071] The following will be based on Figure 1 The scene described by Figure 2 to Figure 6 The product recommendation method of the disclosed embodiment is described in detail.

[0072] Figure 2 The flowchart of the product recommendation method according to the embodiment of the present disclosure is schematically shown.

[0073] like Figure 2 As shown, the method includes operations S210 to S250.

[0074] In operation S210 , in response to receiving a product recommendation request for a target object, a plurality of object information of the target object is determined from an object information library.

[0075] According to an embodiment of the present disclosure, the object information may be basic information of the target object, information on products held by the target object, historical transaction information of the target object, location of the target object, asset status, income status, and the like.

[0076] According to the embodiments of the present disclosure, the object information in the object information library is collected with the authorization of the object. For example, before obtaining the information of the object, a request to obtain the object information can be sent to the object. When the object agrees or authorizes to obtain the user information, the object information is obtained and stored in the object information library.

[0077] According to an embodiment of the present disclosure, the target object may be a customer of the enterprise in other businesses, for example, a customer in a payroll payment business.

[0078] According to an embodiment of the present disclosure, based on the object number included in the product recommendation request for the target object, multiple pieces of object information of the target object can be quickly determined from the object information library.

[0079] According to the embodiments of the present disclosure, the output form of the object information is not limited, and the object information may be output in the form of a data wide table.

[0080] According to the embodiments of the present disclosure, by determining multiple object information, it is possible to profile the target object and obtain portrait information. Based on the portrait information, a shallow analysis of the target object's preferred products can be achieved, thereby helping to recommend the target object's favorite products.

[0081] In operation S220, feature extraction and feature screening are performed on the plurality of object information to obtain object feature sets corresponding to the plurality of candidate products, wherein the object feature set includes a plurality of object features.

[0082] According to an embodiment of the present disclosure, a candidate product is a product that may be recommended to a target object.

[0083] According to the embodiments of the present disclosure, by dividing a plurality of object information into object information corresponding to a plurality of candidate products, the object information related to each candidate product is distinguished, so that the object features can be obtained more accurately and quickly.

[0084] According to the embodiments of the present disclosure, by performing feature extraction and feature screening on the divided object information, the object features corresponding to each candidate product are made to be object features with a high correlation with each candidate product, thereby reducing the redundancy and data volume of the object features and speeding up the process of determining the score value of the target object for each candidate product.

[0085] In operation S230, the object feature sets corresponding to the plurality of candidate products are respectively input into the corresponding prediction models to obtain the score values ​​of the target object for each of the plurality of candidate products.

[0086] According to the embodiments of the present disclosure, each candidate product has its own corresponding prediction model. By inputting the object features corresponding to each candidate product into the corresponding prediction model, the score value of the candidate product can be determined more quickly and accurately.

[0087] In operation S240, based on the score values ​​of the target object for each of the plurality of candidate products and the score values ​​of the plurality of preset objects for each of the plurality of candidate products, similar objects to the target object are determined from the plurality of preset objects.

[0088] According to an embodiment of the present disclosure, the scoring values ​​of the plurality of preset objects for the plurality of candidate products may be obtained based on the following operations.

[0089] Determine multiple preset object information of each of the multiple preset objects from the object information library; for each preset object, perform feature extraction and feature screening on the multiple preset object information respectively, and obtain preset object feature sets corresponding to multiple candidate products respectively; input the object feature sets corresponding to the multiple candidate products respectively into the corresponding prediction models respectively, and obtain the score value of the preset object for the multiple candidate products respectively; since there are multiple preset objects, therefore, obtain the score value of the multiple preset objects for the multiple candidate products respectively.

[0090] According to an embodiment of the present disclosure, in the case where the target object is multiple target objects, the score values ​​of the multiple preset objects for the multiple candidate products can also be obtained based on the following operations.

[0091] The score values ​​of the multiple target objects for the multiple candidate products are used as the score values ​​of the multiple preset objects for the multiple candidate products. When determining similar objects of a certain target object, the similar objects of the target object are calculated based on the score values ​​of the remaining target objects for the multiple candidate products, that is, the target objects other than the target object can be used as preset objects.

[0092] According to an embodiment of the present disclosure, similarity calculations can be performed on the score values ​​of the target object for each candidate product and the score values ​​of the preset object for each of the multiple candidate products, respectively. The similarity between the target object and the preset object is obtained by utilizing the score values ​​of the target object for each candidate product and the score values ​​of each preset object for each candidate product, thereby obtaining similar objects with the same product preference as the target object.

[0093] In operation S250 , a product to be recommended for the target object is determined based on the rating values ​​of the similar objects for the plurality of candidate products.

[0094] According to an embodiment of the present disclosure, the recommendation degree of each candidate product is determined by referring to the score value of each candidate product by similar objects, and the candidate products whose recommendation degree is higher than the recommendation degree threshold or higher than the order threshold after recommendation degree is sorted are selected as products to be recommended.

[0095] According to the product recommendation method provided by the present disclosure, multiple object information of the target object is determined from the object information library, and feature extraction and feature screening are performed on the multiple object information respectively, so as to obtain the object feature set corresponding to each candidate product. The object feature set is input into the corresponding prediction model to obtain the score value of the target object for each of the multiple candidate products, and the similarity is calculated between the score value of the target object for each of the multiple candidate products and the score value of multiple preset objects for each of the multiple candidate products, so as to determine the similar objects of the target object, and starting from the score value of the similar objects for each candidate product, determine the product to be recommended corresponding to the target object. Since each candidate product has a corresponding prediction model, the object feature sets corresponding to each candidate product are input into the corresponding prediction models to more accurately determine the target object's respective ratings for multiple candidate products, that is, the target object's preference for multiple candidate objects. Based on the respective ratings of the multiple candidate products and the similarity of the respective ratings of the multiple preset objects, similar objects are determined for the target object, and the products to be recommended are determined based on the preferences of the similar objects for each candidate product. Therefore, the problem that the recommended products in the related technology cannot meet the user's needs is at least partially solved, and the technical effect of more accurately recommending products to users and improving the user's shopping experience is achieved.

[0096] Figure 3 The flowchart of determining a scoring value according to an embodiment of the present disclosure is schematically shown.

[0097] like Figure 3 As shown, determining the score value includes operations S231 to S232.

[0098] In operation S231, object feature sets corresponding to the plurality of candidate products are input into the corresponding prediction models to obtain a probability value of the target object holding each candidate product.

[0099] In operation S232, the probability value of the target object holding each candidate product is normalized to obtain the target object's respective scoring values ​​for the multiple candidate products.

[0100] According to the embodiments of the present disclosure, the prediction model is not limited and may be any model that can predict the probability value of a target object holding a candidate product, such as a machine learning model, specifically, a convolutional neural network, etc.

[0101] According to an embodiment of the present disclosure, the prediction model is a trained prediction model, wherein the training process of the prediction model may include the following operations.

[0102] Determine the object identification of multiple sample objects; determine the object information corresponding to the multiple object identifications from the object information library; based on the preset field, detect whether the sample object holds the candidate product from the object information of each sample object to obtain the detection result; fuse the object information of the sample object and the detection result to obtain the target information; divide the target information of each sample object according to the candidate products to obtain the target information corresponding to each candidate product. Perform feature extraction and feature screening on the multiple target information to obtain sample feature sets corresponding to the multiple candidate products; input the multiple sample feature sets into the corresponding prediction models to realize the training of multiple prediction models.

[0103] According to an embodiment of the present disclosure, the detection result includes whether the sample object holds the candidate product, and the detection result is equivalent to label data.

[0104] According to an embodiment of the present disclosure, based on a preset field, the object information of each sample object is used to detect whether the sample object holds a candidate product to obtain a detection result. The specific process may be as shown in Table 1: Based on the preset field, the identification of whether the sample object holds a candidate product is detected from the object information of the sample object. The detection result may be represented by a specific code. If the detection result is that the candidate product is held, it is represented by 1. Otherwise, if the sample object does not hold the candidate product, it is represented by 0.

[0105] According to the embodiments of the present disclosure, the number and types of candidate products in Table 1 are only illustrative. In different application scenarios, there may be different numbers and types of candidate products. Similarly, determining the representation and preset fields is also illustrative.

[0106] Table 1

[0107]

[0108] According to the embodiments of the present disclosure, the normalization method is not limited and may be any method that can normalize the probability value of the target object holding each candidate product, for example: linearly transforming each score value so that the probability value is mapped to the interval [0, 100] to obtain each score value, or normalizing the probability value by adopting a normal distribution.

[0109] According to an embodiment of the present disclosure, by using prediction models corresponding to candidate products to predict the probability that the target object holds the candidate, and then normalizing to obtain the target object's score for the candidate product, the target object's score for the candidate product can be obtained more accurately and quickly.

[0110] Figure 4The figure schematically shows a schematic diagram of determining a scoring value based on a prediction model according to an embodiment of the present disclosure.

[0111] like Figure 4 As shown, by inputting sample feature sets corresponding to multiple candidate products of the sample object, such as the sample feature set corresponding to candidate product 1, the sample feature set 402 corresponding to candidate product 2, and the sample feature set 403 corresponding to candidate product n, into untrained prediction models for training, such as the untrained prediction model 404 corresponding to candidate product 1, the untrained prediction model 405 corresponding to candidate product 2, and the untrained prediction model 406 corresponding to candidate product n: trained prediction models are obtained, such as the trained prediction model 407 corresponding to candidate product 1, the trained prediction model 408 corresponding to candidate product 2, and the trained prediction model 409 corresponding to candidate product 1. The trained prediction model 409 corresponding to product n inputs the object feature sets of the target object and each candidate product, such as: the object feature set 410 corresponding to candidate product 1, the object feature set 411 corresponding to candidate product 2, and the object feature set 412 corresponding to candidate product n, into the corresponding prediction models, and obtains the holding probability of the target object for each candidate product, such as the holding probability 413 corresponding to candidate product 1, the holding probability 414 corresponding to candidate product 2, and the holding probability 415 corresponding to candidate product n. The holding probability of the target object for each candidate product is normalized to obtain the target object's score value 416 for each of the multiple candidate products.

[0112] According to an embodiment of the present disclosure, determining similar objects of the target object from multiple preset objects based on the target object's respective ratings of multiple candidate products and the respective ratings of multiple preset objects to the multiple candidate products may include the following operations.

[0113] The similarity between the score value of the target object for each of the multiple candidate products and the score value of each preset object for each of the multiple candidate products is calculated to obtain the similarity between the target object and each preset object; based on the similarity, similar objects of the target object are determined from the multiple preset objects.

[0114] According to the embodiments of the present disclosure, there is no limitation on the similarity calculation method, which may be a cosine similarity calculation method or a Manhattan distance similarity calculation method, wherein the similarity calculation method may be as shown in formula (1).

[0115]

[0116] Among them, sim() is the sub-similarity between a target object and a preset object on a candidate product, i represents the score value of the target object for the candidate product, and j represents the score value of the preset object for the candidate product.

[0117] According to an embodiment of the present disclosure, a similarity value may be obtained by performing a preset calculation on a plurality of sub-similarity values, wherein the preset calculation may be to obtain a similarity value between a target object and a preset object by averaging or summing the values.

[0118] According to an embodiment of the present disclosure, by calculating the similarity between the respective scoring values ​​of each preset object for multiple candidate products and the respective scoring values ​​of the target object for multiple candidate products, the preset objects that have the same product preference for the target object can be determined, wherein, when performing the similarity calculation, the scoring values ​​of some candidate products can be selected for similarity calculation, for example, candidate products whose scoring values ​​of the target object are greater than a scoring threshold are selected for similarity calculation.

[0119] According to an embodiment of the present disclosure, the rating values ​​of multiple preset objects for multiple candidate products can be stored through an object-rating matrix. The object-rating matrix can be stored in an external or internal storage device. Every time a preset time period task is triggered, the object-rating matrix is ​​updated. The preset time period of the preset time period task is not limited and can be one month, one week, etc. The object-rating matrix can be as shown in Table 2.

[0120] Table 2

[0121]

[0122] According to an embodiment of the present disclosure, calculating the similarity between the score value of the target object for each of the candidate products and the score value of each preset object for each of the candidate products to obtain the similarity between the target object and each preset object may include the following operations.

[0123] Determine a first target scoring value greater than a scoring threshold from the respective scoring values ​​of the target object for multiple candidate products; determine a second target scoring value of the target candidate product by each preset object based on the product identifier of the target candidate product corresponding to the first target scoring value; and calculate the similarity between the first target scoring value and the second target scoring value of the target candidate product by each preset object based on the product identifier to obtain the similarity between the target object and each preset object.

[0124] According to the embodiments of the present disclosure, there is no limitation on the scoring threshold, and different scoring thresholds can be set according to actual conditions. For example, the scoring threshold can be zero.

[0125] According to an embodiment of the present disclosure, a target candidate product whose score is greater than a score threshold is selected from the score values ​​of the target object for each of a plurality of candidate products, and the similarity between the score value of the target candidate product and the score value of the target candidate product for each preset object is calculated. For example: the target object has a score of 20 for target candidate product 1 and a score of 80 for target candidate product 2. A preset object has a score of 20 for target candidate product 1 and a score of 40 for target candidate product 2. It can be seen that the similarity between the target object and the preset object is 50%.

[0126] According to an embodiment of the present disclosure, by screening out the candidate products whose ratings of the target object for multiple candidate products are less than or equal to a rating threshold, and then calculating the similarity between each target object and a preset object, it is possible to avoid the problem that the target object has never purchased or does not have object information related to these products, thereby failing to make a correct rating for the candidate products and causing errors in the rating values.

[0127] According to an embodiment of the present disclosure, determining a product to be recommended for a target object based on respective rating values ​​of a plurality of candidate products by similar objects in a similar object set may include the following operations.

[0128] The score values ​​of similar objects for each candidate product are preset to obtain the recommendation degree of each candidate product; the recommendation degree of each candidate product is sorted to obtain a sorting result; based on the sorting result, the candidate products corresponding to the recommendation degrees less than or equal to the preset sequence threshold are used as the products to be recommended for the target object.

[0129] According to the embodiments of the present disclosure, there may be multiple similar objects, and the preset processing method is not limited. The score values ​​of each candidate product in different similar objects may be averaged. For example, similar object 1 scores each candidate product as 10, 20, and 70, and similar object 2 scores each candidate product as 50, 20, and 30, respectively. Then the recommendation degrees of each candidate product are 30, 20, and 50, respectively.

[0130] According to an embodiment of the present disclosure, the preset processing may also be to increase the weight of the score value of each recommended product of the similar objects with high similarity based on the similarities between different similar objects and the target object.

[0131] According to an embodiment of the present disclosure, in the above example, the rankings of the candidate products from high to low are candidate product 3 with a recommendation degree of 50, candidate product 1 with a recommendation degree of 30, and candidate product 2 with a recommendation degree of 20.

[0132] According to the embodiments of the present disclosure, there is no limitation on the preset order threshold, and different preset order thresholds can be set according to different situations. For example, when the preset order threshold is 1, the product to be recommended in the above example is candidate product 3.

[0133] According to the embodiments of the present disclosure, the recommended products for the target object are determined by the rating values ​​of each candidate product given by similar objects, that is, products that may be purchased by objects with similar preferences are used to recommend products to the target object, which can improve the accuracy of product recommendations and better meet user needs.

[0134] According to an embodiment of the present disclosure, feature extraction and feature screening are performed on multiple object information to obtain object feature sets corresponding to multiple candidate products respectively, which may include the following operations.

[0135] Based on the information identifiers of each of the multiple object information, the multiple object information is divided into multiple target object information corresponding to the multiple candidate products respectively; feature extraction is performed on the multiple target object information corresponding to the multiple candidate products respectively to obtain multiple initial feature sets; feature screening is performed on the multiple initial feature sets respectively to obtain object feature sets corresponding to the multiple candidate products respectively.

[0136] According to an embodiment of the present disclosure, information identifiers corresponding to multiple candidate products can be preset, and the information identifiers of multiple object information can be matched with the preset information identifiers, so as to divide the multiple object information into multiple target object information corresponding to the respective objects of the multiple candidate products.

[0137] According to the embodiments of the present disclosure, the manner of extracting features from information of multiple target objects is not limited, and different feature extraction methods may be selected according to different requirements, such as feature mapping and the like.

[0138] According to the embodiments of the present disclosure, there is no limitation on the method of performing feature screening on the initial feature set, and different feature screening methods can be selected according to actual situations. For example, different feature screening methods are adopted based on different feature attributes. More specifically, feature screening can be performed using methods such as Pearson correlation coefficient and chi-square verification.

[0139] According to the embodiments of the present disclosure, the plurality of object information is not specifically limited, and the possible object information extracted in different scenes is different.

[0140] According to an embodiment of the present disclosure, an initial feature set includes multiple initial features; and feature screening is performed on the multiple initial feature sets to obtain object feature sets corresponding to multiple candidate products, which may include the following operations.

[0141] For the initial feature set corresponding to each candidate product, multiple first intermediate features and multiple second intermediate features are determined based on the attribute information of the initial features; variance calculation is performed on the multiple first intermediate features to obtain the variance values ​​of each of the multiple first intermediate features; the first intermediate features whose variance values ​​are less than or equal to the variance threshold are filtered out to obtain the first target feature set; the correlation degree is calculated between the multiple second intermediate features and the candidate products to obtain the correlation degree of each of the multiple second target initial features; the second intermediate features whose correlation degree is less than or equal to the correlation threshold are filtered out to obtain the second target feature set; based on the first target feature set and the second target feature set corresponding to each of the multiple candidate products, object feature sets corresponding to the multiple candidate products are obtained.

[0142] According to an embodiment of the present disclosure, the attribute information may be a type of an initial feature, such as a numerical type or a text type, and the like.

[0143] According to an embodiment of the present disclosure, initial features in an initial feature set are classified into a plurality of first intermediate features and a plurality of second intermediate features based on attribute information.

[0144] According to an embodiment of the present disclosure, since the first intermediate feature is a feature of the data type, the variance of each of the multiple first intermediate features can be determined based on the multiple first intermediate features, thereby determining the degree of discreteness of each first intermediate feature, and screening out the first intermediate features that are less than the variance threshold, thereby filtering out overly concentrated features, thereby ensuring the diversity of the first intermediate features.

[0145] According to an embodiment of the present disclosure, since the second intermediate feature is a text-type feature, the correlation between each second intermediate feature and the candidate product can be calculated, and the features with a low correlation with the candidate product can be filtered out, wherein the method for calculating the correlation between the second intermediate feature and the candidate product is not limited, and the correlation between the second intermediate feature and the candidate product can be calculated by using the Pearson correlation coefficient, Chi-square verification, etc.

[0146] According to an embodiment of the present disclosure, an object feature set of the candidate product is constructed based on the first target feature set and the second target feature set after feature filtering.

[0147] According to the embodiments of the present disclosure, by using different feature screening methods to screen the initial features with different attribute information, the object features corresponding to each candidate product can be obtained more accurately, thereby making the target object's score for each candidate product more accurate.

[0148] Figure 5 A schematic diagram of a product recommendation method according to an embodiment of the present disclosure is schematically shown.

[0149] like Figure 5 As shown, an object identifier 501 of a target object is determined from a product recommendation request, multiple object information 503 is determined from an object information library 502 based on the object identifier, features are extracted and screened for the multiple object information to obtain object feature sets 504 corresponding to multiple candidate products, the object feature sets corresponding to the multiple candidate products are respectively input into prediction models 505 corresponding to the multiple candidate products, and scores 506 of the target object for the multiple candidate products are obtained. Similarity is calculated between the scores 506 of the target object for the multiple candidate products and the scores of multiple preset objects for the multiple candidate products to obtain similar objects 507, and based on the scores of the similar objects for the multiple candidate products, products to be recommended 508 are determined.

[0150] Figure 6 The flowchart of a product recommendation method according to another embodiment of the present disclosure is schematically shown.

[0151] like Figure 6 As shown, the method includes operations S610 to S690.

[0152] In operation S610 , in response to receiving a product recommendation request for a target object, a plurality of object information of the target object is determined from an object information library.

[0153] In operation S620, product information respectively corresponding to a plurality of candidate products is obtained for the plurality of object information.

[0154] According to an embodiment of the present disclosure, the product information may be information about products for which the target object has given positive feedback, such as products that have been positively evaluated, liked, or collected.

[0155] In operation S630, feature extraction and feature screening are performed on the plurality of object information to obtain object feature sets corresponding to the plurality of candidate products respectively.

[0156] In operation S640, feature extraction and feature screening are performed on the product information to obtain target product features corresponding to the positive feedback product.

[0157] In operation S650, feature extraction is performed on candidate product information of each of the plurality of candidate products to obtain candidate product features corresponding to each of the plurality of candidate products.

[0158] In operation S660, the object feature sets corresponding to the plurality of candidate products are respectively input into the corresponding prediction models to obtain the score values ​​of the target object for the plurality of candidate products.

[0159] In operation S670, similarities are calculated between the target product feature and the candidate product features corresponding to each of the plurality of candidate products, to determine a plurality of first products to be recommended for the target object.

[0160] According to an embodiment of the present disclosure, by calculating the first similarity between the target product features and the candidate product features of each candidate product, the products that the target object may be interested in are obtained, and the obtained products corresponding to the first similarities with each candidate object are sorted to obtain a sorting result, and the candidate products in the sorting result that are greater than the first sorting threshold are used as the first products to be recommended.

[0161] In operation S680, a plurality of second to-be-recommended products for the target object are determined based on the respective score values ​​of the target object for the plurality of candidate products and the respective score values ​​of the plurality of preset objects for the plurality of candidate products.

[0162] According to an embodiment of the present disclosure, similar objects to the target object are determined from multiple preset objects by comparing the score values ​​of the target object for each of the multiple candidate products and the score values ​​of multiple preset objects for each of the multiple candidate products; and a second product to be recommended for the target object is determined based on the score values ​​of the similar objects for the multiple candidate products.

[0163] In operation S690, based on the plurality of first products to be recommended and the plurality of second products to be recommended, products to be recommended for the target object are obtained.

[0164] According to an embodiment of the present disclosure, a plurality of first products to be recommended and a plurality of second products to be recommended are intersected to obtain products to be recommended for a target object.

[0165] Based on the above product recommendation method, the present disclosure also provides a product recommendation device. Figure 7 The device is described in detail.

[0166] Figure 7 The structural block diagram of the product recommendation device according to an embodiment of the present disclosure is schematically shown.

[0167] like Figure 7 As shown, the product recommendation device 700 of this embodiment includes an object information determination module 710 , an object feature set determination module 720 , a score value determination module 730 , a similar object determination module 740 and a to-be-recommended product determination module 750 .

[0168] The object information determination module 710 is configured to determine a plurality of object information of the target object from an object information library in response to receiving a product recommendation request for the target object.

[0169] The object feature set determination module 720 is used to extract and screen features of multiple object information to obtain object feature sets corresponding to multiple candidate products, wherein the object feature set includes multiple object features.

[0170] The score value determination module 730 is used to input the object feature sets corresponding to the multiple candidate products into the corresponding prediction models to obtain the score values ​​of the target object for each of the multiple candidate products.

[0171] The similar object determination module 740 is used to determine similar objects of the target object from multiple preset objects based on the respective scoring values ​​of the target object to the multiple candidate products and the respective scoring values ​​of the multiple preset objects to the multiple candidate products.

[0172] The to-be-recommended product determination module 750 is used to determine the to-be-recommended product for the target object based on the rating values ​​of the similar objects to the multiple candidate products.

[0173] According to an embodiment of the present disclosure, the score value determination module 730 includes: a probability value prediction submodule and a score value determination submodule.

[0174] The probability value prediction submodule is used to input the object feature sets corresponding to the multiple candidate products into the corresponding prediction models to obtain the probability value of the target object holding each candidate product.

[0175] The scoring value determination submodule is used to normalize the probability value of the target object holding each candidate product to obtain the target object's respective scoring values ​​for multiple candidate products.

[0176] According to an embodiment of the present disclosure, the similar object determination module 740 includes: a similarity determination submodule and a similar object determination submodule.

[0177] The similarity determination submodule is used to calculate the similarity between the score value of the target object for each of the multiple candidate products and the score value of each preset object for each of the multiple candidate products, so as to obtain the similarity between the target object and each preset object.

[0178] The similar object determination submodule is used to determine similar objects of the target object from a plurality of preset objects based on similarity.

[0179] According to an embodiment of the present disclosure, the similarity determination submodule includes: a first score value determination unit, a second score value determination unit, and a similarity determination unit.

[0180] The first rating value determination unit is used to determine a first target rating value greater than a rating threshold from the rating values ​​of the target object for each of the plurality of candidate products.

[0181] The second scoring value determining unit is used to determine the second target scoring value of each preset object for the target candidate product based on the product identification of the target candidate product corresponding to the first target scoring value.

[0182] The similarity determination unit is used to calculate the similarity between the first target rating value and the second target rating value of each preset object for the target candidate product based on the product identification, so as to obtain the similarity between the target object and each preset object.

[0183] According to an embodiment of the present disclosure, the to-be-recommended product determination module 750 includes: a recommendation degree determination submodule, a ranking submodule and a to-be-recommended product determination submodule.

[0184] The recommendation degree determination submodule is used to preset the score value of each candidate product by similar objects to obtain the recommendation degree of each candidate product.

[0185] The sorting submodule is used to sort the recommendation degree of each candidate product to obtain a sorting result.

[0186] The to-be-recommended product determination submodule is used to select, based on the sorting results, candidate products corresponding to a recommendation degree that is less than or equal to a preset order threshold as to-be-recommended products for the target object.

[0187] According to an embodiment of the present disclosure, the object feature set determination module 720 includes: an information division submodule, a feature extraction submodule, and a feature screening submodule.

[0188] The information division submodule is used to divide the multiple object information into multiple target object information corresponding to multiple candidate products respectively based on the information identifiers of the multiple object information.

[0189] The feature extraction submodule is used to extract features from multiple target object information corresponding to multiple candidate products respectively to obtain multiple initial feature sets.

[0190] The feature screening submodule is used to screen the multiple initial feature sets respectively to obtain object feature sets corresponding to the multiple candidate products respectively.

[0191] According to an embodiment of the present disclosure, the initial feature set includes a plurality of initial features. The feature screening submodule includes: a feature determination unit, a variance value calculation unit, a first filtering unit, a correlation calculation submodule, a second filtering unit and an object feature set determination submodule.

[0192] The feature determination unit is used to determine, for an initial feature corresponding to each candidate product, a plurality of first intermediate features and a plurality of second intermediate features based on attribute information of the initial feature.

[0193] The variance value calculation unit is used to perform variance calculation on the multiple first intermediate features to obtain the variance values ​​of the multiple first intermediate features.

[0194] The first filtering unit is used to filter out the first intermediate features whose variance values ​​are less than or equal to the variance threshold to obtain a first target feature set.

[0195] The correlation calculation submodule is used to calculate the correlation between the multiple second intermediate features and the candidate products to obtain the correlation of each of the multiple second target initial features.

[0196] The second filtering unit is used to filter out the second intermediate features whose correlation degree is less than or equal to the correlation degree threshold to obtain a second target feature set.

[0197] The object feature set determination submodule is used to obtain object feature sets corresponding to multiple candidate products respectively based on the first target feature sets and the second target feature sets corresponding to the multiple candidate products respectively.

[0198] According to an embodiment of the present disclosure, any multiple modules of the object information determination module 710, the object feature set determination module 720, the score value determination module 730, the similar object determination module 740, and the to-be-recommended product determination module 750 can be combined in one module for implementation, or any one of the modules can be split into multiple modules. Alternatively, at least part of the functions of one or more of these modules can be combined with at least part of the functions of other modules and implemented in one module. According to an embodiment of the present disclosure, at least one of the object information determination module 710, the object feature set determination module 720, the score value determination module 730, the similar object determination module 740, and the to-be-recommended product determination module 750 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application specific integrated circuit (ASIC), or can be implemented by hardware or firmware such as any other reasonable way of integrating or packaging the circuit, or implemented in any one of the three implementation methods of software, hardware, and firmware or in an appropriate combination of any of them. Alternatively, at least one of the object information determination module 710, the object feature set determination module 720, the rating value determination module 730, the similar object determination module 740, and the product to be recommended determination module 750 may be at least partially implemented as a computer program module, which may perform corresponding functions when executed.

[0199] Figure 8 A block diagram of an electronic device suitable for implementing a product recommendation method according to an embodiment of the present disclosure is schematically shown.

[0200] like Figure 8As shown, the electronic device 800 according to an embodiment of the present disclosure includes a processor 801, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 802 or a program loaded from a storage part 808 to a random access memory (RAM) 803. The processor 801 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 801 may also include an onboard memory for caching purposes. The processor 801 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present disclosure.

[0201] In RAM 803, various programs and data required for the operation of electronic device 800 are stored. Processor 801, ROM 802 and RAM 803 are connected to each other via bus 804. Processor 801 performs various operations of the method flow according to the embodiment of the present disclosure by executing the program in ROM 802 and / or RAM 803. It should be noted that the program can also be stored in one or more memories other than ROM 802 and RAM 803. Processor 801 can also perform various operations of the method flow according to the embodiment of the present disclosure by executing the program stored in the one or more memories.

[0202] According to an embodiment of the present disclosure, the electronic device 800 may further include an input / output (I / O) interface 805, which is also connected to the bus 804. The electronic device 800 may further include one or more of the following components connected to the input / output (I / O) interface 805: an input portion 806 including a keyboard, a mouse, etc.; an output portion 807 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage portion 808 including a hard disk, etc.; and a communication portion 809 including a network interface card such as a LAN card, a modem, etc. The communication portion 809 performs communication processing via a network such as the Internet. A drive 810 is also connected to the input / output (I / O) interface 805 as needed. A removable medium 811, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 810 as needed, so that a computer program read therefrom is installed into the storage portion 808 as needed.

[0203] The present disclosure also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or may exist independently without being assembled into the device / apparatus / system. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed, the method according to the embodiment of the present disclosure is implemented.

[0204] According to an embodiment of the present disclosure, the computer-readable storage medium may be a non-volatile computer-readable storage medium, for example, may include but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, an apparatus or a device. For example, according to an embodiment of the present disclosure, a computer-readable storage medium may include the ROM 802 and / or RAM 803 described above and / or one or more memories other than ROM 802 and RAM 803.

[0205] The embodiment of the present disclosure also includes a computer program product, which includes a computer program, and the computer program contains program code for executing the method shown in the flowchart. When the computer program product is run in a computer system, the program code is used to enable the computer system to implement the product recommendation method provided by the embodiment of the present disclosure.

[0206] The above functions defined in the system / device of the embodiment of the present disclosure are executed when the computer program is executed by the processor 801. According to the embodiment of the present disclosure, the system, device, module, unit, etc. described above can be implemented by a computer program module.

[0207] In one embodiment, the computer program may rely on tangible storage media such as optical storage devices, magnetic storage devices, etc. In another embodiment, the computer program may also be transmitted and distributed in the form of signals on a network medium, and downloaded and installed through the communication part 809, and / or installed from a removable medium 811. The program code contained in the computer program may be transmitted using any appropriate network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0208] In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 809, and / or installed from the removable medium 811. When the computer program is executed by the processor 801, the above functions defined in the system of the embodiment of the present disclosure are performed. According to the embodiment of the present disclosure, the system, device, means, module, unit, etc. described above can be implemented by a computer program module.

[0209] According to an embodiment of the present disclosure, the program code for executing the computer program provided by the embodiment of the present disclosure can be written in any combination of one or more programming languages. Specifically, these computing programs can be implemented using high-level process and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, Java, C++, python, "C" language or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on the remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (for example, using an Internet service provider to connect through the Internet).

[0210] The flow charts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flow chart or block diagram can represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0211] It will be appreciated by those skilled in the art that the features described in the various embodiments and / or claims of the present disclosure may be combined and / or combined in a variety of ways, even if such combinations and / or combinations are not explicitly described in the present disclosure. In particular, the features described in the various embodiments and / or claims of the present disclosure may be combined and / or combined in a variety of ways without departing from the spirit and teachings of the present disclosure. All of these combinations and / or combinations fall within the scope of the present disclosure.

[0212] The embodiments of the present disclosure are described above. However, these embodiments are only for illustrative purposes and are not intended to limit the scope of the present disclosure. Although the embodiments are described above separately, this does not mean that the measures in the various embodiments cannot be used in combination to advantage. The scope of the present disclosure is defined by the attached claims and their equivalents. Without departing from the scope of the present disclosure, those skilled in the art may make a variety of substitutions and modifications, which should all fall within the scope of the present disclosure.

Claims

1. A product recommendation method, comprising: In response to receiving a product recommendation request for a target object, determining a plurality of object information of the target object from an object information library; Extracting and screening features of the plurality of object information to obtain object feature sets corresponding to the plurality of candidate products, wherein the object feature sets include a plurality of object features; Inputting the object feature sets corresponding to the plurality of candidate products into the corresponding prediction models respectively, and obtaining the score values ​​of the target object for the plurality of candidate products respectively; Based on the score values ​​of the target object for each of the plurality of candidate products and the score values ​​of the plurality of preset objects for each of the plurality of candidate products, determining a similar object to the target object from the plurality of preset objects; Based on the ratings of the similar objects for the plurality of candidate products, a product to be recommended for the target object is determined.

2. The method according to claim 1, wherein: The step of inputting the object feature sets corresponding to the plurality of candidate products into the prediction models corresponding to the respective ones to obtain the score values ​​of the target object for the plurality of candidate products respectively comprises: Inputting object feature sets corresponding to the plurality of candidate products into the corresponding prediction models to obtain a probability value of the target object holding each of the candidate products; The probability value of the target object holding each of the candidate products is normalized to obtain the score value of the target object for each of the candidate products.

3. The method according to claim 1, wherein: The determining of similar objects to the target object from a plurality of preset objects based on the respective scoring values ​​of the target object for the plurality of candidate products and the respective scoring values ​​of the plurality of preset objects for the plurality of candidate products comprises: Calculate the similarity between the score value of the target object for each of the candidate products and the score value of each of the preset objects for each of the candidate products, to obtain the similarity between the target object and each of the preset objects; Based on the similarity, a similar object to the target object is determined from the plurality of preset objects.

4. The method according to claim 3, wherein: The calculating the similarity between the score value of the target object for each of the candidate products and the score value of each of the preset objects for each of the candidate products to obtain the similarity between the target object and each of the preset objects includes: Determine a first target score value greater than a score threshold from the score values ​​of the target object for each of the plurality of candidate products; Determining, based on the product identifier of the target candidate product corresponding to the first target rating value, a second target rating value of the target candidate product by each of the preset objects; Based on the product identification, similarity calculation is performed between the first target rating value and the second target rating value of each preset object for the target candidate product to obtain the similarity between the target object and each preset object.

5. The method according to claim 1, wherein: The determining of the product to be recommended for the target object based on the score values ​​of the plurality of candidate products respectively given by the similar objects in the similar object set includes: Preset the score of each candidate product by the similar object to obtain the recommendation degree of each candidate product; Ranking the recommendation degree of each of the candidate products to obtain a ranking result; Based on the ranking result, the candidate products corresponding to the recommendation degrees less than or equal to the preset order threshold are used as the products to be recommended for the target object.

6. The method according to claim 1, wherein: The feature extraction and feature screening of the plurality of object information to obtain object feature sets corresponding to the plurality of candidate products respectively includes: Based on the information identifiers of the plurality of object information, the plurality of object information is divided into a plurality of target object information corresponding to each of the plurality of candidate products; Extracting features of a plurality of target object information corresponding to each of the plurality of candidate products to obtain a plurality of initial feature sets; Feature screening is performed on the multiple initial feature sets respectively to obtain object feature sets corresponding to the multiple candidate products respectively.

7. The method according to claim 1, wherein: The initial feature set includes a plurality of initial features; The step of performing feature screening on the plurality of initial feature sets to obtain object feature sets corresponding to the plurality of candidate products includes: For an initial feature set corresponding to each of the candidate products, determining a plurality of first intermediate features and a plurality of second intermediate features based on attribute information of the initial features; performing variance calculation on the plurality of first intermediate features to obtain variance values ​​of the plurality of first intermediate features; Filter out the first intermediate features whose variance values ​​are less than or equal to the variance threshold to obtain a first target feature set; Calculating the correlation between the plurality of the second intermediate features and the candidate products to obtain the correlation between the plurality of the second target initial features; Filter out the second intermediate features whose correlation degree is less than or equal to the correlation degree threshold to obtain a second target feature set; Based on the first target feature set and the second target feature set corresponding to each of the plurality of candidate products, the object feature sets corresponding to the plurality of candidate products are obtained.

8. A product recommendation device, comprising: an object information determination module, configured to determine a plurality of object information of the target object from an object information library in response to receiving a product recommendation request for the target object; An object feature set determination module, used for performing feature extraction and feature screening on the plurality of object information to obtain object feature sets corresponding to the plurality of candidate products, wherein the object feature set includes a plurality of object features; A scoring value determination module, used to input the object feature sets corresponding to the plurality of candidate products into the corresponding prediction models, and obtain the scoring values ​​of the target object for the plurality of candidate products; A similar object determination module, configured to determine similar objects of the target object from a plurality of preset objects based on the respective score values ​​of the target object for the plurality of candidate products and the respective score values ​​of the plurality of preset objects for the plurality of candidate products; The module for determining products to be recommended is used to determine products to be recommended for the target object based on the rating values ​​of the similar objects for the plurality of candidate products.

9. An electronic device, comprising: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors are enabled to execute the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, causes the processor to execute the method according to any one of claims 1 to 7.

Images