CN113393306A

CN113393306A - Product recommendation method and device, electronic equipment and computer readable medium

Info

Publication number: CN113393306A
Application number: CN202110778954.5A
Authority: CN
Inventors: 李昭莹; 傅强; 张舜华; 冷真敏; 李娟�
Original assignee: China Construction Bank Corp
Current assignee: China Construction Bank Corp
Priority date: 2021-07-09
Filing date: 2021-07-09
Publication date: 2021-09-14

Abstract

The invention discloses a product recommendation method and device, electronic equipment and a computer readable medium, and relates to the technical field of artificial intelligence. One embodiment of the method comprises: acquiring user payment information, and determining a project label of a target payment project according to a project label added to at least one payment project in advance; the user payment information comprises a target payment item and payment time; calculating preference weight of a user on a target payment item according to the payment time, and taking an item label and the preference weight of the target payment item as user preference attributes; calculating the preference degree of the user to the candidate products according to the product labels, the label weights and the user preference attributes of the candidate products in the candidate product set; and selecting a recommended product from the candidate product set according to the preference degree, and outputting the recommended product. The method and the system can automatically match the recommended products without manual processing, and the recommendation accuracy is high.

Description

Product recommendation method and device, electronic equipment and computer readable medium

Technical Field

The invention relates to the technical field of artificial intelligence, in particular to a product recommendation method, a product recommendation device, electronic equipment and a computer readable medium.

Background

With the development of financial institutions, financial products are continuously enriched, and more choices are provided for customers. Conventional financial products rely on manual recommendations, which are labor intensive and time consuming, by the financial service provider analyzing the customer's needs and determining the appropriate financial product to recommend to the customer.

Disclosure of Invention

In view of this, embodiments of the present invention provide a product recommendation method, an apparatus, an electronic device, and a computer-readable medium, where the method determines a preference attribute of a user by using an association between payment information of the user and payment items, and further automatically matches a recommended product by using an association between the payment items and candidate products, so that manual processing is not required, and recommendation accuracy is high.

To achieve the above object, according to an aspect of an embodiment of the present invention, there is provided a product recommendation method.

The product recommendation method provided by the embodiment of the invention comprises the following steps: acquiring user payment information, and determining a project label of a target payment project according to a project label added to at least one payment project in advance; the user payment information comprises the target payment item and payment time; calculating preference weight of a user on the target payment item according to the payment time, and taking an item tag of the target payment item and the preference weight as user preference attributes; calculating the preference degree of the user for the candidate products according to the product labels, label weights and the user preference attributes of the candidate products in the candidate product set; and selecting a recommended product from the candidate product set according to the preference degree, and outputting the recommended product.

Optionally, the calculating, according to the product labels, the label weights, and the user preference attributes of the candidate products in the candidate product set, a preference degree of the user for the candidate products includes: according to the item labels of the user preference attributes, searching candidate products of which the product labels are associated with the item labels from the candidate product set; decomposing the item label into a product of the product label of the candidate product and the corresponding label weight to obtain a decomposition result; and multiplying the decomposition result by the preference weight of the user preference attribute to obtain the preference degree of the user to the candidate product.

Optionally, the searching for the candidate product with the product tag associated with the item tag from the candidate product set according to the item tag of the user preference attribute includes: and acquiring candidate products containing the item labels in the product labels in an inverted index mode according to the item labels of the user preference attributes.

Optionally, calculating a preference weight of the user for the target payment item according to the payment time includes: calculating the preference weight of the user on the target payment item by using a pre-constructed preference weight calculation formula; wherein the preference weight calculation formula is as follows:

w_n＝w_n-1×exp(-α×t)

in the formula, w_nThe nth preference weight; w is a_n-1Preference weight for (n-1) < th >; alpha is a cooling coefficient; t is the interval time; i is a positive integer, and the maximum value is (current time-payment time)/t.

Optionally, the method further comprises: extracting set keywords from the product information of the candidate product, and calculating the word frequency and the inverse document frequency of the keywords; and determining the product label and the label weight of the candidate product according to the word frequency and the inverse document frequency.

Optionally, the determining the product label and the label weight of the candidate product according to the word frequency and the inverse document frequency includes: calculating the product of the word frequency and the inverse document frequency to obtain a TF-IDF value of the keyword; according to the TF-IDF value, performing descending sorting on the keywords, and taking the keywords with a first quantity set in the top sorting as product labels of the candidate products; and normalizing the TF-IDF value, and taking the normalized TF-IDF value as the label weight of the product label.

Optionally, the selecting a recommended product from the candidate product set according to the preference degree includes: and according to the preference degree, the candidate products are arranged in a descending order, and the candidate products with the second quantity which are arranged in the front order are taken as recommended products.

Optionally, the method further comprises: and acquiring basic user information, and constructing a user portrait according to the basic user information and the user preference attribute.

Optionally, the constructing a user representation according to the user basic information and the user preference attribute includes: extracting user basic attributes from the user basic information to generate an initial user portrait; updating the initial user representation using the user preference attribute to obtain a final user representation.

Optionally, the method further comprises: and screening the original product set according to the product admission threshold and the user portrait to obtain a candidate product set.

Optionally, the method further comprises: respectively carrying out item marking on the at least one payment item to obtain corresponding label attributes; and solving a union set of the category attribute of the at least one payment item and the corresponding label attribute to obtain an item label of the at least one payment item.

Optionally, the respectively performing item labeling on the at least one payment item to obtain corresponding tag attributes includes: inputting the project text of the at least one payment project into a pre-trained project marking model for label prediction to obtain a label attribute corresponding to the project text; and the item labeling model is obtained by training a machine model by using a labeled sample text set.

Optionally, the acquiring the user payment information includes: and acquiring user payment information generated by the payment operation under the condition of detecting the payment operation.

To achieve the above object, according to another aspect of the embodiments of the present invention, there is provided a product recommendation device.

The product recommendation device of the embodiment of the invention comprises: the tag determining module is used for acquiring user payment information and determining a project tag of a target payment project according to a project tag added to at least one payment project in advance; the user payment information comprises the target payment item and payment time; the first calculation module is used for calculating the preference weight of the user on the target payment item according to the payment time, and taking the item label of the target payment item and the preference weight as the preference attribute of the user; the second calculation module is used for calculating the preference degree of the user on the candidate product according to the product label, the label weight and the user preference attribute of the candidate product in the candidate product set; and the product recommending module is used for selecting recommended products from the candidate product set according to the preference degree and outputting the recommended products.

Optionally, the second computing module is further configured to search, according to the item tag of the user preference attribute, a candidate product whose product tag is associated with the item tag from the candidate product set; decomposing the item label into a product of the product label of the candidate product and the corresponding label weight to obtain a decomposition result; and multiplying the decomposition result by the preference weight of the user preference attribute to obtain the preference degree of the user to the candidate product.

Optionally, the second computing module is further configured to obtain, by using an inverted index manner, a candidate product including the item tag in the product tag according to the item tag of the user preference attribute.

Optionally, the first calculating module is further configured to calculate a preference weight of the user for the target payment item by using a pre-constructed preference weight calculation formula; wherein the preference weight calculation formula is as follows:

w_n＝w_n-1×exp(-α×t)

in the formula, w_nIs the nth timeA preference weight; w is a_n-1Preference weight for (n-1) < th >; alpha is a cooling coefficient; t is the interval time; i is a positive integer, and the maximum value is (current time-payment time)/t.

Optionally, the apparatus further comprises: the product labeling module is used for extracting set keywords from the product information of the candidate product and calculating the word frequency and the inverse document frequency of the keywords; and determining the product label and the label weight of the candidate product according to the word frequency and the inverse document frequency.

Optionally, the product labeling module is further configured to calculate a product of the word frequency and the inverse document frequency to obtain a TF-IDF value of the keyword; according to the TF-IDF value, performing descending sorting on the keywords, and taking the keywords with a first quantity set in the top sorting as product labels of the candidate products; and normalizing the TF-IDF value, and taking the normalized TF-IDF value as the label weight of the product label.

Optionally, the product recommending module is further configured to sort the candidate products in a descending order according to the preference degree, and use the candidate products with the second number set in the top order as recommended products.

Optionally, the apparatus further comprises: and the construction module is used for acquiring the basic information of the user and constructing the user portrait according to the basic information of the user and the preference attribute of the user.

Optionally, the building module is further configured to extract a user basic attribute from the user basic information, and generate an initial user representation; updating the initial user representation using the user preference attribute to obtain a final user representation.

Optionally, the apparatus further comprises: and the screening module is used for screening the original product set according to the product admission threshold and the user portrait to obtain a candidate product set.

Optionally, the apparatus further comprises: the item label determining module is used for respectively carrying out item labeling on the at least one payment item to obtain corresponding label attributes; and solving a union set of the category attribute of the at least one payment item and the corresponding label attribute to obtain an item label of the at least one payment item.

Optionally, the item label determining module is further configured to input the item text of the at least one payment item to a pre-trained item labeling model for label prediction, so as to obtain a label attribute corresponding to the item text; and the item labeling model is obtained by training a machine model by using a labeled sample text set.

Optionally, the tag determination module is further configured to acquire user payment information generated by the payment operation when the payment operation is detected.

To achieve the above object, according to still another aspect of embodiments of the present invention, there is provided an electronic apparatus.

An electronic device of an embodiment of the present invention includes: one or more processors; a storage device for storing one or more programs which, when executed by the one or more processors, cause the one or more processors to implement a method for recommending a product according to an embodiment of the present invention.

To achieve the above object, according to still another aspect of embodiments of the present invention, there is provided a computer-readable medium.

A computer-readable medium of an embodiment of the present invention has a computer program stored thereon, which when executed by a processor implements a product recommendation method of an embodiment of the present invention.

One embodiment of the above invention has the following advantages or benefits: the user preference attribute is determined by using the association between the user payment information and the payment items, and then the recommended product is automatically matched by using the association between the payment items and the candidate product, so that manual processing is not needed, and the recommendation accuracy is high.

Further effects of the above-mentioned non-conventional alternatives will be described below in connection with the embodiments.

Drawings

The drawings are included to provide a better understanding of the invention and are not to be construed as unduly limiting the invention. Wherein:

FIG. 1 is a schematic diagram of the main steps of a product recommendation method according to an embodiment of the invention;

FIG. 2 is a schematic diagram of a main flow of a product recommendation method according to an embodiment of the invention;

FIG. 3 is a flow diagram illustrating an implementation of updating an initial user representation according to an embodiment of the present invention;

FIG. 4 is a schematic diagram of the major modules of a product recommendation device according to an embodiment of the present invention;

FIG. 5 is an exemplary system architecture diagram in which embodiments of the present invention may be employed;

FIG. 6 is a schematic block diagram of a computer system suitable for use with the electronic device to implement an embodiment of the invention.

Detailed Description

Exemplary embodiments of the present invention are described below with reference to the accompanying drawings, in which various details of embodiments of the invention are included to assist understanding, and which are to be considered as merely exemplary. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the invention. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.

Terms related to the present embodiment are explained below.

User portrait: is a virtual representation of a real user, is a user model built on top of a series of real data. A tagged user model is abstracted according to information such as social attributes, living habits, consumption behaviors and the like of a user.

TF-IDF: the method is a common weighting technology for information retrieval and data mining, and is used for evaluating the importance degree of a word to a file set or a corpus. TF is Term Frequency (Term Frequency) and IDF is Inverse text Frequency (Inverse Document Frequency).

Inverted indexing: the reason for this is that in practical applications it is necessary to look up records based on the values of attributes. Each entry in such an index table includes an attribute value and the address of the record having the attribute value. Since the attribute value is not determined by the record but the position of the record is determined by the attribute value, it is called an inverted index.

Fig. 1 is a schematic diagram of the main steps of a product recommendation method according to an embodiment of the present invention. As shown in fig. 1, the product recommendation method according to the embodiment of the present invention is implemented by a product recommendation device, and mainly includes the following steps:

step S101: the method comprises the steps of obtaining user payment information, and determining a project label of a target payment project according to a project label added for at least one payment project in advance. After the user carries out payment operation on the payment platform, the payment platform can generate user payment information. In an embodiment, the user payment information may include a target payment item, payment time, payment amount, payment method, and the like.

Adding project labels for at least one payment project contained in a payment platform in advance, and storing the project names and the corresponding project labels of the payment projects in a database. After the product recommendation device acquires user payment information from the payment platform, item labels of payment items with the same item names as the target payment items are searched from the database, and the searched item labels are used as item labels of the target payment items.

Step S102: and calculating the preference weight of the user on the target payment item according to the payment time, and taking the item label of the target payment item and the preference weight as the preference attribute of the user. User preferences are a dynamic feature that describes the user's preferences for selecting payment items. The user preference is time sensitive and the longer the time, the lower the user preference and therefore the preference weight needs to decay over time.

In the embodiment, the preference weight of the user on the target payment item is calculated through a pre-constructed preference weight calculation formula (the preference weight calculation formula can enable the preference weight to decay with time), and then the user preference attribute can be described in a mode of combining the item label and the preference weight.

Step S103: and calculating the preference degree of the user for the candidate products according to the product labels, the label weights and the user preference attributes of the candidate products in the candidate product set. The step is used for searching candidate products from the candidate product set according to the preference attribute of the user and calculating the preference degree of the user on the candidate products. The product tag of the candidate product is associated with the item tag of the user preference attribute, for example, if the product tag of the candidate product includes an item tag of at least one user preference attribute, the two may be considered to be associated.

Specifically, according to item labels of user preference attributes, candidate products of which the product labels are associated with the item labels are searched from a candidate product set; then decomposing the item label into the product of the product label of the candidate product and the corresponding label weight to obtain a decomposition result; and multiplying the decomposition result by the preference weight of the preference attribute of the user to obtain the preference degree of the user to the candidate product.

Step S104: and selecting a recommended product from the candidate product set according to the preference degree, and outputting the recommended product. And selecting the candidate product with high preference degree from the candidate product set as the recommended product according to the preference degree, and outputting the recommended product. The candidate product selected here may be a candidate product whose preference degree is greater than a set threshold value, or may be the top N candidate products whose preference degree is high. Wherein N is an integer.

Fig. 2 is a schematic diagram of a main flow of a product recommendation method according to an embodiment of the present invention. As shown in fig. 2, the product recommendation method according to the embodiment of the present invention mainly includes the following steps:

step S201: and acquiring basic information of the user and constructing an initial user portrait. When a user registers on a payment platform, some basic information, such as a mobile phone number, an identity card number and the like, needs to be entered. The payment platform will register basic information, and at the same time, the relevant interface can be called to further inquire other basic information, such as credit investigation, risk rating, etc. The basic information entered by the user and the inquired basic information jointly form the basic information of the user.

And extracting the basic user attributes from the basic user information to generate an initial user portrait. The basic attribute of the user is a static feature, and can comprise gender, age, occupation, marital status, education level, city and the like.

Step S202: and acquiring user payment information, analyzing the user payment information, updating the initial user portrait and obtaining a final user portrait. When a user logs in a payment platform to perform specific transactions, such as payment operation, the payment platform records information of target payment items, payment time, payment amount, payment mode, user account, user asset rating and the like, and stores the information in a database.

The user can depict the preference attribute of the user and can reflect the dynamic change of the preference of the user in the payment operation of the target payment item, so that the product recommendation device can acquire the information from the database under the condition of detecting the payment operation so as to further perfect the initial user portrait, ensure the dynamic update of the user portrait and improve the subsequent recommendation effect.

In an embodiment, the user representation includes primarily user base attributes and user preference attributes. The user preference attribute is obtained by analyzing the user payment information, and includes an item tag and a preference weight of the target payment item, and the specific implementation is described with reference to fig. 3. It can be understood that, in order to ensure the accuracy of the user portrait, the user portrait can be constructed by performing preprocessing such as data cleaning and standardization on the user basic information and the user payment information, and then analyzing and extracting the preprocessed user basic information and the preprocessed user payment information.

Step S203: and acquiring product information of the candidate products in the candidate product set, and adding product labels and label weights to the candidate products. Product information for the candidate product may come from a third party organization, also stored in the database. In an embodiment, the product information may include information such as a product code, a product name, a manager, a docket facility, a product status, a risk rating, a basic introduction, and the like.

Extracting set keywords from product information of candidate products, and calculating the word frequency and the inverse document frequency of each keyword by using a TF-IDF (word frequency-inverse document frequency) algorithm; and then determining the product label and the label weight of the candidate product according to the word frequency and the inverse document frequency. The calculation formula of the word frequency of each keyword is as follows:

in the formula, TF_i,jRepresenting the frequency of occurrence of the keyword i in the product information of the candidate product j; n is_i,jRepresenting the number of times of occurrence of the keyword i in the product information of the candidate product j; sigma_kn_k,jAnd k is the number of the keywords, and represents the sum of the occurrence times of all the keywords in the product information of the candidate product j.

The calculation formula of the inverse document frequency of each keyword is as follows:

in the formula, IDF_iRepresenting the inverse document frequency of the keyword i, and measuring the importance of the keyword; y represents the total number of product information in the corpus; y is_iIndicating the amount of product information containing the keyword i.

When the product label and the label weight of the candidate product are determined, the product of the word frequency and the inverse document frequency can be calculated to obtain the TF-IDF value of the keyword; then, according to the TF-IDF value, performing descending sorting on the plurality of keywords, and taking the keywords with the first quantity set in the top sorting as product labels of the candidate products; and carrying out normalization processing on the TF-IDF value, and taking the normalized TF-IDF value as the label weight of the product label. Wherein, the calculation formula of the TF-IDF value of each keyword is as follows:

TF-IDF_i,j＝TF_i,j×IDF_iequation 3

Step S204: and calculating the preference degree of the user for the candidate product according to the product label, the label weight and the user portrait of the candidate product. Step S202 obtains item labels and preference weights of the target payment items, step S203 obtains product labels and label weights of the candidate products, and calculates the preference degree of the user for the candidate products using the following preference degree calculation formula:

wherein, U is the preference degree of the user to the candidate product, d is the number of item labels, and S_xIndicates the preference weight, T, corresponding to the x-th item tag_xDenotes the x-th item tag, O_xpP-th product label, w, representing a candidate product associated with the x-th item label_xpIndicating the label weight corresponding to the p-th product label.

In an embodiment, the candidate product associated with the item tag refers to a candidate product in which at least one item tag is included in the product tag. For example, if the product tag of a candidate product is identical to the item tag, the candidate product is a candidate product associated with the item tag. For another example, if a product tag of a candidate product includes multiple item tags, the candidate product may also be considered as a candidate product associated with the item tags.

In equation 4, the 2 nd equal sign is used to assign T_xIs decomposed into w_x*O_xAnd acquiring a candidate product containing at least one item label in the product label according to the item label of the user preference attribute in an inverted index mode. Wherein, w_xProduct tag, O, representing a candidate product associated with the xth item tag_xIndicating the corresponding weight of the product label.

Considering the candidate products as the basis of the vector space in equation 4, different item tags may be associated to the same candidate product. The same items need to be merged on the right side of the last equal sign in the formula 4, and the label weights in front of the same candidate products are added, so that the preference degree of the user on the candidate products can be obtained.

Step S205: and selecting a recommended product from the candidate product set according to the preference degree, and outputting the recommended product. In an embodiment, the candidate products may be sorted in a descending order according to the preference degree, and the candidate products with a second number (for example, the top N) set in the top order may be recommended to the user as recommended products.

Some financial products have product admission thresholds, such as user age, credit, deposit amount, loan amount, etc., so that the candidate products of the candidate product set need to meet the corresponding product admission thresholds. In a preferred embodiment, the candidate product set may be obtained by: and screening the original product set according to the product admission threshold and the user portrait to obtain a candidate product set. Wherein the original product set includes all financial products that are planned for recommendation. The above processing makes the present embodiment obtain candidate products that meet the admission threshold.

Specifically, for each financial product in the original product set, whether the user portrait meets a product access threshold of the current financial product can be compared, and if the user portrait meets the product access threshold, the current financial product is reserved; and if the product admission threshold is not met, deleting the current financial product. After each financial product in the original product set is compared in the above manner, the remaining financial products in the original product set constitute a candidate product set.

According to the embodiment, the tag labels are added to the payment items and the candidate products by utilizing the similarity between the payment items and the candidate products, the user images are built in real time according to the payment operation of the user on a payment platform, then the candidate products of the candidate product set are screened according to the user images, the recommended products are automatically matched in real time and are recommended to the user in a personalized mode, the recommendation accuracy is high, the recommendation service can be continuously provided in all weather, and the real-time performance is good. The recommended products can be embedded into the advertising positions of the user interface of the payment platform and recommended to the user, the user does not need any operation, and user experience is improved.

FIG. 3 is a flow chart illustrating an implementation of updating an initial user representation according to an embodiment of the invention. As shown in FIG. 3, the implementation process of updating an initial user portrait according to the embodiment of the present invention includes the following steps:

step S301: and determining the project label of the target payment project according to the project label added to at least one payment project in advance. Adding project labels for at least one payment project in a payment platform in advance, and storing the project names and the corresponding project labels of the payment projects in a database. After the user payment information is acquired, the item label of the payment item with the same item name as the target payment item can be searched from the database, and the searched item label is used as the item label of the target payment item.

For example, if the project name of the target payment project is "south power grid payment", the project label of the project name of the payment project of "south power grid payment" is searched from the database, and the searched project label is used as the project label of the target payment project.

The specific implementation of adding a project tag to at least one payment project in the payment platform may be: respectively carrying out item marking on at least one payment item to obtain corresponding label attributes; and then, integrating the category attribute of at least one payment item with the corresponding label attribute to obtain an item label of at least one payment item.

The item labeling is to label some characteristics of the payment item, such as government affairs, military camp, temple, house lease and the like, and the labeling result is the label attribute of the payment item. A payment item may have a plurality of tag attributes. The category attribute can comprise a large category attribute and a small category attribute to which the payment item belongs.

In the embodiment, when item marking is performed, item text of at least one payment item can be input to a pre-trained item labeling model for label prediction, so that a label attribute corresponding to the item text is obtained. The project labeling model is obtained by training a machine model by using a labeled sample text set. The item text may be item information of the payment item, such as an item name, item details, and the like. And the sample text in the sample text set is the item information of the payment item marked with the label attribute. The process realizes automatic project labeling without manual processing.

In the embodiment, it is assumed that the payment platform has hundreds of thousands of payment items, including payment items of southern power grid, power supply bureau of guangzhou, school and miscellaneous fees, school and campus card recharge, and the like. These payment items can be divided into 8 major categories, each of which is subdivided into several sub-categories, such as:

paying and recharging: electricity, water, property, etc.;

automobile service: traffic fines, parking fees, etc.;

travel for business: train tickets, bus tickets, sight spot tickets, etc.;

catering and entertainment: movie tickets, lottery tickets, etc.;

and (3) education service: miscellaneous fee learning, kindergarten payment, training fee and the like;

medical insurance: social security, banking services, etc.;

paying the tax: non-tax payment, national tax, and land tax;

more services: the cost of the trade and the trade.

In the above example, each payment item is associated with at least one subclass. And respectively solving and collecting the major attribute, the minor attribute and the label attribute of each payment item to obtain the item label of each payment item.

Step S302: and calculating the preference weight of the user to the target payment item according to the payment time in the user payment information, and taking the combination of the item label and the preference weight of the target payment item as the user preference attribute. The user preference attributes are depicted in a combination of item labels and preference weights. When a user selects a payment item to conduct transaction, an item label corresponding to the payment item can be marked for the user, and initial preference weight is given.

Since the user preference is time-sensitive, the preference weight needs to decay over time. In an embodiment, the above attenuation requirement may be realized by constructing a preference weight calculation formula through an exponential function. Specifically, the preference weight calculation formula is:

w_n＝w_n-1xexp (- α × t) formula 5

In the formula, w_nThe nth preference weight; w is a_n-1Preference weight for (n-1) < th >; alpha is a cooling coefficient; t is the interval time; i is a positive integer, and the maximum value is (current time-payment time)M)/t.

In specific application, the cooling coefficient can be customized. The cooling coefficient takes a smaller value if it is desired to favor a slower decay in weight, and a larger value otherwise. For example, the initial preference weight of the user for the first payment operation on a certain payment item may be defined as 1, and the preference weight is expected to be cooled to 0.5 after 1 day, then the cooling coefficient may be obtained from equation 1.

Step S303: the initial user representation is updated with user preference attributes to obtain a final user representation. The initial user representation is refined using the user preference attributes to obtain a final user representation.

FIG. 4 is a schematic diagram of the main modules of a product recommendation device according to an embodiment of the present invention. As shown in fig. 4, the product recommendation device 400 according to the embodiment of the present invention mainly includes:

the tag determination module 401 is configured to acquire user payment information, and determine a project tag of a target payment project according to a project tag added to at least one payment project in advance. After the user carries out payment operation on the payment platform, the payment platform can generate user payment information. In an embodiment, the user payment information may include a target payment item, payment time, payment amount, payment method, and the like.

A first calculating module 402, configured to calculate a preference weight of the user for the target payment item according to the payment time, and use an item tag of the target payment item and the preference weight as a user preference attribute. User preferences are a dynamic feature that describes the user's preferences for selecting payment items. The user preference is time sensitive and the longer the time, the lower the user preference and therefore the preference weight needs to decay over time.

A second calculating module 403, configured to calculate, according to the product label, the label weight, and the user preference attribute of a candidate product in the candidate product set, a preference degree of the user for the candidate product. The module is used for searching candidate products from the candidate product set according to the preference attribute of the user and calculating the preference degree of the user on the candidate products. The product tag of the candidate product is associated with the item tag of the user preference attribute, for example, if the product tag of the candidate product includes an item tag of at least one user preference attribute, the two may be considered to be associated.

And a product recommending module 404, configured to select a recommended product from the candidate product set according to the preference degree, and output the recommended product. And selecting the candidate product with high preference degree from the candidate product set as the recommended product according to the preference degree, and outputting the recommended product. The candidate product selected here may be a candidate product whose preference degree is greater than a set threshold value, or may be the top N candidate products whose preference degree is high. Wherein N is an integer.

In addition, the product recommendation 400 according to the embodiment of the present invention may further include: a product labeling module, a building module, a screening module, and an item tag determination module (not shown in FIG. 4). The product labeling module is used for extracting set keywords from the product information of the candidate product and calculating the word frequency and the inverse document frequency of the keywords; and determining the product label and the label weight of the candidate product according to the word frequency and the inverse document frequency. And the construction module is used for acquiring the basic information of the user and constructing the user portrait according to the basic information of the user and the preference attribute of the user.

And the screening module is used for screening the original product set according to the product admission threshold and the user portrait to obtain a candidate product set. The item label determining module is used for respectively carrying out item labeling on the at least one payment item to obtain corresponding label attributes; and solving a union set of the category attribute of the at least one payment item and the corresponding label attribute to obtain an item label of the at least one payment item.

From the above description, it can be seen that the user preference attribute is determined by using the association between the user payment information and the payment item, and then the recommended product is automatically matched by using the association between the payment item and the candidate product, so that manual processing is not required, and the recommendation accuracy is high.

Fig. 5 illustrates an exemplary system architecture 500 of a product recommendation method or device to which embodiments of the invention may be applied.

As shown in fig. 5, the system architecture 500 may include

terminal devices

501, 502, 503, a network 504, and a server 505. The network 504 serves to provide a medium for communication links between the

terminal devices

501, 502, 503 and the server 505. Network 504 may include various connection types, such as wired, wireless communication links, or fiber optic cables, to name a few.

The user may use the

terminal devices

501, 502, 503 to interact with a server 505 over a network 504 to receive or send messages or the like. The

terminal devices

501, 502, 503 may have various communication client applications installed thereon, such as a shopping application, a web browser application, a search application, an instant messaging tool, a mailbox client, social platform software, and the like.

The

terminal devices

501, 502, 503 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop portable computers, desktop computers, and the like.

The server 505 may be a server that provides various services, such as a background management server that supports user payment information generated by the user using the

terminal devices

501, 502, and 503. The background management server can determine item labels and preference weights of target payment items, calculate preference degrees of the users on candidate products, filter recommended products and the like, and feed back processing results (such as recommended products) to the terminal equipment.

It should be noted that the product recommendation method provided by the embodiment of the present invention is generally executed by the server 505, and accordingly, the product recommendation apparatus is generally disposed in the server 505.

It should be understood that the number of terminal devices, networks, and servers in fig. 5 is merely illustrative. There may be any number of terminal devices, networks, and servers, as desired for implementation.

The invention also provides an electronic device and a computer readable medium according to the embodiment of the invention.

The electronic device of the present invention includes: one or more processors; a storage device for storing one or more programs which, when executed by the one or more processors, cause the one or more processors to implement a method for recommending a product according to an embodiment of the present invention.

The computer-readable medium of the present invention has stored thereon a computer program which, when executed by a processor, implements a product recommendation method of an embodiment of the present invention.

Referring now to FIG. 6, shown is a block diagram of a computer system 600 suitable for use with the electronic device implementing an embodiment of the present invention. The terminal device shown in fig. 6 is only an example, and should not bring any limitation to the functions and the scope of use of the embodiments of the present invention.

As shown in fig. 6, the computer system 600 includes a Central Processing Unit (CPU)601 that can perform various appropriate actions and processes according to a program stored in a Read Only Memory (ROM)602 or a program loaded from a storage section 608 into a Random Access Memory (RAM) 603. In the RAM 603, various programs and data necessary for the operation of the system 600 are also stored. The CPU 601, ROM 602, and RAM 603 are connected to each other via a bus 604. An input/output (I/O) interface 605 is also connected to bus 604.

The following components are connected to the I/O interface 605: an input portion 606 including a keyboard, a mouse, and the like; an output portion 607 including a display such as a Cathode Ray Tube (CRT), a Liquid Crystal Display (LCD), and the like, and a speaker; a storage section 608 including a hard disk and the like; and a communication section 609 including a network interface card such as a LAN card, a modem, or the like. The communication section 609 performs communication processing via a network such as the internet. The driver 610 is also connected to the I/O interface 605 as needed. A removable medium 611 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like is mounted on the drive 610 as necessary, so that a computer program read out therefrom is mounted in the storage section 608 as necessary.

In particular, according to the embodiments of the present disclosure, the processes described above with reference to the flowcharts may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program embodied on a computer readable medium, the computer program comprising program code for performing the method illustrated in the flow chart. In such an embodiment, the computer program may be downloaded and installed from a network through the communication section 609, and/or installed from the removable medium 611. The computer program performs the above-described functions defined in the system of the present invention when executed by the Central Processing Unit (CPU) 601.

It should be noted that the computer readable medium shown in the present invention can be a computer readable signal medium or a computer readable storage medium or any combination of the two. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples of the computer readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the present invention, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In the present invention, however, a computer readable signal medium may include a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, fiber optic cable, RF, etc., or any suitable combination of the foregoing.

The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams or flowchart illustration, and combinations of blocks in the block diagrams or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

The modules described in the embodiments of the present invention may be implemented by software or hardware. The described modules may also be provided in a processor, which may be described as: a processor includes a tag determination module, a first computation module, a second computation module, and a product recommendation module. The names of the modules do not form a limitation on the modules themselves under certain conditions, for example, the tag determination module may also be described as a module for acquiring user payment information and determining a project tag of a target payment project according to a project tag added to at least one payment project in advance.

As another aspect, the present invention also provides a computer-readable medium that may be contained in the apparatus described in the above embodiments; or may be separate and not incorporated into the device. The computer readable medium carries one or more programs which, when executed by a device, cause the device to comprise: acquiring user payment information, and determining a project label of a target payment project according to a project label added to at least one payment project in advance; the user payment information comprises the target payment item and payment time; calculating preference weight of a user on the target payment item according to the payment time, and taking an item tag of the target payment item and the preference weight as user preference attributes; calculating the preference degree of the user for the candidate products according to the product labels, label weights and the user preference attributes of the candidate products in the candidate product set; and selecting a recommended product from the candidate product set according to the preference degree, and outputting the recommended product.

According to the technical scheme of the embodiment of the invention, the user preference attribute is determined by utilizing the association between the user payment information and the payment item, and then the recommended product is automatically matched by utilizing the association between the payment item and the candidate product, so that manual processing is not needed, and the recommendation accuracy is high.

The above-described embodiments should not be construed as limiting the scope of the invention. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions can occur, depending on design requirements and other factors. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for recommending products, comprising:

acquiring user payment information, and determining a project label of a target payment project according to a project label added to at least one payment project in advance; the user payment information comprises the target payment item and payment time;

calculating preference weight of a user on the target payment item according to the payment time, and taking an item tag of the target payment item and the preference weight as user preference attributes;

calculating the preference degree of the user for the candidate products according to the product labels, label weights and the user preference attributes of the candidate products in the candidate product set;

and selecting a recommended product from the candidate product set according to the preference degree, and outputting the recommended product.

2. The method of claim 1, wherein calculating the preference degree of the user for the candidate product according to the product label, the label weight and the user preference attribute of the candidate product in the candidate product set comprises:

according to the item labels of the user preference attributes, searching candidate products of which the product labels are associated with the item labels from the candidate product set;

decomposing the item label into a product of the product label of the candidate product and the corresponding label weight to obtain a decomposition result;

and multiplying the decomposition result by the preference weight of the user preference attribute to obtain the preference degree of the user to the candidate product.

3. The method of claim 2, wherein the finding a candidate product from the set of candidate products whose product tag is associated with the item tag according to the item tag of the user preference attribute comprises:

and acquiring candidate products containing the item labels in the product labels in an inverted index mode according to the item labels of the user preference attributes.

4. The method of claim 1, wherein calculating the preference weight of the user for the target payment item according to the payment time comprises:

calculating the preference weight of the user on the target payment item by using a pre-constructed preference weight calculation formula; wherein the preference weight calculation formula is as follows:

w_n＝w_n-1×exp(-α×t)

5. The method of claim 1, further comprising:

extracting set keywords from the product information of the candidate product, and calculating the word frequency and the inverse document frequency of the keywords;

and determining the product label and the label weight of the candidate product according to the word frequency and the inverse document frequency.

6. The method of claim 5, wherein determining the product label and label weight of the candidate product based on the word frequency and the inverse document frequency comprises:

calculating the product of the word frequency and the inverse document frequency to obtain a TF-IDF value of the keyword;

according to the TF-IDF value, performing descending sorting on the keywords, and taking the keywords with a first quantity set in the top sorting as product labels of the candidate products;

and normalizing the TF-IDF value, and taking the normalized TF-IDF value as the label weight of the product label.

7. The method of claim 1, wherein selecting a recommended product from the set of candidate products according to the preference level comprises:

and according to the preference degree, the candidate products are arranged in a descending order, and the candidate products with the second quantity which are arranged in the front order are taken as recommended products.

8. The method of claim 1, further comprising:

and acquiring basic user information, and constructing a user portrait according to the basic user information and the user preference attribute.

9. The method of claim 8, wherein constructing a user representation based on the user base information and the user preference attributes comprises:

extracting user basic attributes from the user basic information to generate an initial user portrait;

updating the initial user representation using the user preference attribute to obtain a final user representation.

10. The method of claim 8, further comprising:

and screening the original product set according to the product admission threshold and the user portrait to obtain a candidate product set.

11. The method of claim 1, further comprising:

respectively carrying out item marking on the at least one payment item to obtain corresponding label attributes;

and solving a union set of the category attribute of the at least one payment item and the corresponding label attribute to obtain an item label of the at least one payment item.

12. The method of claim 11, wherein the labeling the at least one payment item for the item to obtain the corresponding label attribute comprises:

inputting the project text of the at least one payment project into a pre-trained project marking model for label prediction to obtain a label attribute corresponding to the project text; and the item labeling model is obtained by training a machine model by using a labeled sample text set.

13. The method according to any one of claims 1 to 12, wherein the obtaining user payment information comprises:

and acquiring user payment information generated by the payment operation under the condition of detecting the payment operation.

14. A product recommendation device, comprising:

the tag determining module is used for acquiring user payment information and determining a project tag of a target payment project according to a project tag added to at least one payment project in advance; the user payment information comprises the target payment item and payment time;

the first calculation module is used for calculating the preference weight of the user on the target payment item according to the payment time, and taking the item label of the target payment item and the preference weight as the preference attribute of the user;

the second calculation module is used for calculating the preference degree of the user on the candidate product according to the product label, the label weight and the user preference attribute of the candidate product in the candidate product set;

and the product recommending module is used for selecting recommended products from the candidate product set according to the preference degree and outputting the recommended products.

15. An electronic device, comprising:

one or more processors;

a storage device for storing one or more programs,

when executed by the one or more processors, cause the one or more processors to implement the method of any one of claims 1-13.

16. A computer-readable medium, on which a computer program is stored, which, when being executed by a processor, carries out the method according to any one of claims 1-13.