CN111966886A - Object recommendation method, object recommendation device, electronic equipment and storage medium - Google Patents

Object recommendation method, object recommendation device, electronic equipment and storage medium Download PDF

Info

Publication number
CN111966886A
CN111966886A CN201910420177.XA CN201910420177A CN111966886A CN 111966886 A CN111966886 A CN 111966886A CN 201910420177 A CN201910420177 A CN 201910420177A CN 111966886 A CN111966886 A CN 111966886A
Authority
CN
China
Prior art keywords
class
objects
feature
data
feature data
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201910420177.XA
Other languages
Chinese (zh)
Inventor
李文静
王晶
黄�俊
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Jingdong Century Trading Co Ltd
Beijing Jingdong Shangke Information Technology Co Ltd
Original Assignee
Beijing Jingdong Century Trading Co Ltd
Beijing Jingdong Shangke Information Technology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Beijing Jingdong Century Trading Co Ltd, Beijing Jingdong Shangke Information Technology Co Ltd filed Critical Beijing Jingdong Century Trading Co Ltd
Priority to CN201910420177.XA priority Critical patent/CN111966886A/en
Publication of CN111966886A publication Critical patent/CN111966886A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9536—Search customisation based on social or collaborative filtering
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9535—Search customisation based on user profiles and personalisation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/06—Buying, selling or leasing transactions
    • G06Q30/0601—Electronic shopping [e-shopping]
    • G06Q30/0631—Item recommendations

Landscapes

  • Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Accounting & Taxation (AREA)
  • Finance (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Economics (AREA)
  • General Business, Economics & Management (AREA)
  • Strategic Management (AREA)
  • Marketing (AREA)
  • Development Economics (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The disclosure provides an object recommendation method, an object recommendation device, an electronic device and a computer readable storage medium, and belongs to the technical field of computers. The method comprises the following steps: acquiring characteristic data of a plurality of first-class objects; dividing the feature data of each first class object into a plurality of groups of feature data based on the association between the feature data and a plurality of classes of second class objects; converting each group of feature data of each first class object into a feature sub-vector respectively, and determining the feature vector of each first class object according to a plurality of feature sub-vectors of each first class object; determining similarity between feature vectors of the plurality of first class objects; and determining a first class of objects for recommendation according to the similarity. The method and the device can make the recommended object have pertinence, and can improve the object recommendation efficiency and the recommendation accuracy.

Description

Object recommendation method, object recommendation device, electronic equipment and storage medium
Technical Field
The present disclosure relates to the field of computer technologies, and in particular, to an object recommendation method, an object recommendation apparatus, an electronic device, and a computer-readable storage medium.
Background
With the increasingly widespread popularization and application of the internet in various industries, enterprises in various fields such as e-commerce, internet finance, life service, games and the like all strive to collect and analyze data of objects such as commodities, users and the like through the internet to mine demands and effectively recommend the objects, including: recommending users to merchants, recommending merchants or goods to users, recommending other users to users, etc.
Most of the existing object recommendation methods collect massive data of various objects, construct object features based on the data, and determine similar objects as recommendation objects. However, in practical applications, there may be relations between different types of objects, for example, in an e-commerce platform, there is usually a behavioral relationship between a user and a commodity, and it is difficult to express the degree of relation between the user and the commodity by using the above method. In addition, since the links between the objects of different types may not correspond to each other, for example, each user may not have a behavior relationship with all the objects, so that the acquired data has sparsity, and a situation of data missing or data misconcentration may occur. Generally, when the method is used for object recommendation, the acquired data is directly transcoded into data with preset dimensionality, and the influence of the association degree between different objects on data characteristics is ignored, so that the pertinence and the effectiveness are lacked when the recommendation is performed.
The learning model can accurately predict the objects, and effective recommendation can not be performed according to the association degree of different objects, so that the pertinence is lacked.
It is to be noted that the information disclosed in the above background section is only for enhancement of understanding of the background of the present disclosure, and thus may include information that does not constitute prior art known to those of ordinary skill in the art.
Disclosure of Invention
The present disclosure provides an object recommendation method, an object recommendation apparatus, an electronic device, and a computer-readable storage medium, so as to overcome, at least to a certain extent, the problems of low recommendation accuracy and lack of pertinence of the existing object recommendation method.
Additional features and advantages of the disclosure will be set forth in the detailed description which follows, or in part will be obvious from the description, or may be learned by practice of the disclosure.
According to an aspect of the present disclosure, there is provided an object recommendation method including: acquiring characteristic data of a plurality of first-class objects; dividing the feature data of each first class object into a plurality of groups of feature data based on the association between the feature data and a plurality of classes of second class objects; converting each group of feature data of each first class object into a feature sub-vector respectively, and determining the feature vector of each first class object according to a plurality of feature sub-vectors of each first class object; determining similarity between feature vectors of the plurality of first class objects; and determining a first class of objects for recommendation according to the similarity.
In an exemplary embodiment of the present disclosure, the second class object includes N classes; the dividing the feature data of each first class object into a plurality of groups of feature data based on the association between the feature data and a plurality of classes of second class objects comprises: for each feature data, if the feature data is associated with the ith category in the N categories, dividing the feature data into ith groups of feature data so as to divide the feature data of each first-class object into multiple groups of feature data; and the ith group of characteristic data has a corresponding relation with the ith category of the second class object.
In an exemplary embodiment of the present disclosure, the converting each group of feature data of each first class object into a feature sub-vector, and determining the feature vector of each first class object according to a plurality of feature sub-vectors of each first class object includes: acquiring a preset dimension of each group of feature data, and converting each group of feature data into a feature sub-vector according to the preset dimension; and splicing the characteristic sub-vectors of each first class object to obtain the characteristic vector of each first class object.
In an exemplary embodiment of the present disclosure, the determining the first class of objects for recommendation according to the similarity includes: clustering the plurality of first class objects based on the similarity; and determining the first class of objects for recommendation according to the clustering result.
In an exemplary embodiment of the present disclosure, the clustering the plurality of first class objects based on the similarity includes: clustering each first class object for the first time based on the feature vector of each first class object to obtain a plurality of sets of the first class objects; respectively judging whether each set meets a preset condition; and if the set meets the preset condition, clustering the first class of objects in the set again.
In an exemplary embodiment of the present disclosure, the preset condition is that the number of the first class objects in the set is less than a first threshold; if the set meets the preset condition, clustering the first class of objects in the set again comprises: and if the number of the first class objects in the set reaches the first threshold value, determining a hash parameter according to the number of the first class objects in the set, and clustering the first class objects in the set again by adopting a locality sensitive hash algorithm.
In an exemplary embodiment of the present disclosure, the determining the similarity between the feature vectors of the plurality of first class objects includes: acquiring a reference object in the plurality of first-class objects; determining similarity between feature vectors of each first class object and the reference object; the determining the first class of objects for recommendation according to the similarity includes: and determining the first class of objects with the similarity reaching a second threshold value with the feature vector of the reference object as the first class of objects for recommendation.
According to an aspect of the present disclosure, there is provided an object recommending apparatus including: the data acquisition module is used for acquiring characteristic data of a plurality of first-class objects; the data grouping module is used for dividing the characteristic data of each first class object into a plurality of groups of characteristic data based on the association between the characteristic data and a plurality of classes of second class objects; the vector determination module is used for converting each group of feature data of each first class object into feature sub-vectors respectively and determining the feature vector of each first class object according to the plurality of feature sub-vectors of each first class object; the similarity determining module is used for determining the similarity among the feature vectors of the plurality of first-class objects; and the object determining module is used for determining the first class of objects for recommendation according to the similarity.
In an exemplary embodiment of the present disclosure, the second class object includes N classes; the data grouping module is used for dividing the characteristic data into ith group of characteristic data if the characteristic data is associated with ith category in the N categories so as to divide the characteristic data of each first-class object into a plurality of groups of characteristic data; and the ith group of characteristic data has a corresponding relation with the ith category of the second class object.
In an exemplary embodiment of the present disclosure, the vector determination module includes: the dimension acquiring unit is used for acquiring the preset dimension of each group of feature data and converting each group of feature data into feature sub-vectors according to the preset dimension; and the vector splicing unit is used for splicing the characteristic sub-vectors of each first class object to obtain the characteristic vector of each first class object.
In an exemplary embodiment of the present disclosure, the object determination module includes: a clustering unit, configured to cluster the plurality of first class objects based on the similarity; and the determining unit is used for determining the first class of objects for recommendation according to the clustering result.
In an exemplary embodiment of the present disclosure, the clustering unit includes: the pre-clustering subunit is used for performing primary clustering on each first class object based on the feature vector of each first class object to obtain a plurality of sets related to the first class objects; the judging subunit is used for respectively judging whether each set meets a preset condition; and the re-clustering subunit is used for re-clustering the first class of objects in the set if the set meets the preset condition.
In an exemplary embodiment of the present disclosure, the preset condition is that the number of the first class objects in the set is less than a first threshold; and the re-clustering subunit is used for determining a hash parameter according to the number of the first class objects in the set and re-clustering the first class objects in the set by adopting a locality sensitive hash algorithm if the number of the first class objects in the set reaches the first threshold value.
In an exemplary embodiment of the present disclosure, the similarity determination module includes: a reference object acquisition unit configured to acquire a reference object from among the plurality of objects of the first type; the similarity determining unit is used for determining the similarity between the feature vectors of each first class object and the reference object; the object determination module includes: and the threshold judging unit is used for determining the first class of objects with the similarity reaching a second threshold with the feature vector of the reference object as the first class of objects for recommendation.
According to an aspect of the present disclosure, there is provided an electronic device including: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to perform the method of any one of the above via execution of the executable instructions.
According to an aspect of the present disclosure, there is provided a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the method of any one of the above.
Exemplary embodiments of the present disclosure have the following advantageous effects:
the method comprises the steps of obtaining feature data of first-class objects, dividing the feature data of each first-class object into a plurality of groups of feature data based on the association between the first-class object and a plurality of classes of second-class objects, converting the feature data into feature sub-vectors to determine feature vectors, and determining the first-class objects for recommendation according to the similarity between the first objects. On one hand, compared with a general recommendation method in which feature data is directly processed, the exemplary embodiment preprocesses the feature data based on the association between the feature data of the first class of objects and multiple classes of the second class of objects, determines the characteristics of different first class of objects, and makes object recommendation more targeted and effective; on the other hand, the feature data of the first-class object is divided into a plurality of groups of feature data, the feature vectors are determined after the feature sub-vectors are converted, and the feature vectors of the first-class object which are more abundant and effective can be obtained, so that accurate recommendation can be performed.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and together with the description, serve to explain the principles of the disclosure. It is to be understood that the drawings in the following description are merely exemplary of the disclosure, and that other drawings may be derived from those drawings by one of ordinary skill in the art without the exercise of inventive faculty.
FIG. 1 schematically illustrates a flow chart of an object recommendation method in the present exemplary embodiment;
FIG. 2 schematically illustrates a flowchart of another object recommendation method in the present exemplary embodiment;
FIG. 3 schematically illustrates a sub-flowchart of an object recommendation method in the present exemplary embodiment;
FIG. 4 schematically illustrates a sub-flowchart of another object recommendation method in the present exemplary embodiment;
FIG. 5 is a sub-flowchart schematically illustrating yet another object recommendation method in the present exemplary embodiment;
fig. 6 is a block diagram schematically showing the configuration of an object recommending apparatus in the present exemplary embodiment;
fig. 7 schematically illustrates an electronic device for implementing the above method in the present exemplary embodiment;
fig. 8 schematically illustrates a computer-readable storage medium for implementing the above-described method in the present exemplary embodiment.
Detailed Description
Example embodiments will now be described more fully with reference to the accompanying drawings. Example embodiments may, however, be embodied in many different forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the concept of example embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.
An exemplary embodiment of the present disclosure first provides an object recommendation method, in which an object may be an object that users, goods, and the like of platforms such as e-commerce, social, and the like may use for recommendation. The application scenarios of the present exemplary embodiment may be: in the e-commerce platform, the goods are recommended to the user, in the life service platform, the user is recommended to the merchant, in the social network platform, the user (recommending friends, interested people) is recommended to the user, and the like, which is not particularly limited by the disclosure.
Fig. 1 shows the flow steps of the present exemplary embodiment, which may include steps S110 to S150. Each step is described in detail below:
step S110, feature data of a plurality of first-class objects are acquired.
Step S120, based on the association between the above feature data and the multiple categories of the second-class object, dividing the feature data of each first-class object into multiple sets of feature data.
In this exemplary embodiment, the first type of object refers to a recommended object, and the second type of object is another type of object that is associated with the first type of object, for example, in an e-commerce platform, a user may generate behavior data about a commodity, such as browsing behavior, attention behavior, and the like for the commodity, that is, there is a connection between the user and the commodity, when recommending the user to a merchant, the first type of object may be the user, and the second type of object may be the commodity; in recommending items to a user, the first type of object may be an item, the second type of object may be a user, and so on. The feature data may be obtained by counting original data of the first type object, for example, when the first type object is a user, the feature data may be behavior data of the user within a period of time (for example, browsing data, concern data, and purchase adding data (purchase adding means that the user puts a commodity into a shopping cart, and purchase adding data may reflect a purchase intention of the user); or when the first type object is a commodity, the characteristic data can be the browsing amount, the purchase adding amount, the marking amount and the like of a certain commodity in a period of time. In some cases, part or all of the data of the feature data may also be raw data of the first type object, such as the order time of the user, the amount of the deal or the attribute of the goods, the selling price or the selling record, and so on. The present exemplary embodiment may determine the feature data by acquiring data of some or all of the first type of object of a certain platform, for example, if the present exemplary embodiment is applied to a scenario in which a user recommends a commodity in an e-commerce platform, the feature data may be determined by acquiring data of the commodity in a certain time period in a certain e-commerce platform, or may also determine the feature data according to data of the commodity with a higher scoring value by scoring the behavior data of the commodity (for example, browsing 1 score, focusing 2 score, placing 3 scores, and the like). If the method is applied to a scene that the user is recommended to the user in the social platform, all user data in a certain social platform can be obtained to determine the characteristic data.
The categories may reflect the categories of the second-class objects under different index degrees, for example, when the second-class object is a commodity, the commodity may be classified into hot commodity, normal commodity, cold commodity, and the like according to indexes such as browsing amount, attention amount, or purchase amount of the commodity; when the second type object is a user, the user can be classified into an active user, an inactive user and the like according to indexes such as the operation frequency, the login frequency, the click frequency and the like of the user. By classifying the second class of objects, it is advantageous in the present exemplary embodiment to perform differentiated processing on the first class of objects of different classes. Before step S120, the plurality of categories of the second-class object may be determined by the data of the first-class object in step S110, or by acquiring other data. Based on the association between the feature data and the plurality of classes of the second class of objects, it can be determined which feature data in the first class of objects belongs to which class of the second class of objects, so that the feature data corresponding to the first class of objects of that class can be divided into a set of feature data.
In an exemplary embodiment, the second class of objects includes N classes. Step S120 may include: for each feature data, if the feature data is associated with the ith category in the N categories, the feature data is divided into ith group of feature data, so that the feature data of each first-class object is divided into multiple groups of feature data, wherein i is any natural number less than or equal to N, and the ith group of feature data has a corresponding relation with the ith category of the second-class object.
In the present exemplary embodiment, the number of categories of the second class object may be N, and when the second class object is a commodity, in addition to the categories of hot commodities, normal commodities, cold commodities, and the like, the commodities may be classified more finely according to the degree of hot commodities, such as first-level hot commodities, second-level hot commodities, and the like. If the feature data of the first class of objects is associated with a class of the second class of objects, for example, a certain shopping behavior of the user is a shopping behavior of a hot commodity, the feature data of the shopping behavior of the user can be divided into a group, and the group can be used for storing the feature data of the hot commodity. The exemplary embodiment is directed to generalizing the feature data of the first class object in the same second class object category, so that the feature data can be processed differently according to different types, and in combination with the association between the first class object and the second class object category, the pertinence of recommendation of the first class object is enhanced.
For example, when the first type object is a user, the characteristic data of the user a is shown in the following table:
TABLE 1
Number of times of browsing Number of additional purchases Number of comments Number of orders Degree of attention
First class of merchandise 10 0 5 2 3
Second kind of commodities 80 50 3 10 15
Class III articles 20 10 0 5 20
The second category of commodity is a hot commodity, the third category of commodity is a normal commodity, the first category of commodity is a cold commodity, and according to the association between the feature data and the multiple categories of the second category of object, the feature data of the user a can be divided into: first group of feature data (feature data on popular goods) "browsing times 30, buying times 50, commenting times 3, placing times 10, attention degree 15"; a second group of feature data (feature data on normal goods) "browsing times 20, purchase adding times 10, review times 0, order placing times 5, attention degree 20"; the third group of feature data (feature data on cold goods) "number of views 10, number of purchases 0, number of comments 5, number of orders placed 2, degree of attention 3".
In an exemplary embodiment, feature data may also be present that is not associated with each of the N categories, and these feature data may be divided into a particular set of feature data.
In practical applications, not all the feature data are related to the category of the second class object, and some data may appear, which is not related to determining the recommended object, such as login time, purchase time, and user ID (identity) of the user, so that these data may be divided into a specific set of feature data, and whether to process the set of feature data is determined according to needs.
Step S130, converting each group of feature data of each first-class object into a feature sub-vector, and determining the feature vector of each first-class object according to the plurality of feature sub-vectors of each first-class object.
In the exemplary embodiment, for convenience of computational analysis, each set of feature data of each first-class object may be converted into a feature sub-vector, respectively, since the feature data may be in a non-numerical form, such as a comment, a preference degree, or a commodity type of a user on a commodity. Thus, the feature data may be numerically converted to generate feature subvectors. Determining the feature vector of each first class object in different modes according to each generated feature sub-vector, specifically, splicing different feature sub-vectors to determine the feature vector; or by calculating the values of the same dimension in each feature sub-vector, such as adding, arithmetic mean, weighted average, etc., to obtain the feature vector, etc., the determination method of the feature vector of the first kind of object may be various, and the disclosure does not particularly limit this.
It should be noted that the dimensions of the feature sub-vectors of each set of feature data conversion may be the same or different, and in an exemplary embodiment, the step S130 may include the following steps:
acquiring a preset dimension of each group of feature data, and converting each group of feature data into a feature sub-vector according to the preset dimension;
and splicing the characteristic sub-vectors of each first class object to obtain the characteristic vector of each first class object.
The preset dimension may be used to reflect a data attribute of each group of feature data, for example, when the first class object is a user and the second class object is a commodity, the preset dimension may include behavior data of the user about the commodity, such as browsing times, purchase adding times, comment times, order placing times, or attention degree. The preset dimensions of different sets of feature data may be the same or different. For example, regarding the user a, the feature sub-vectors of the hot goods, the normal goods, and the cold goods may be set to be a preset dimension (e.g., 10) with the same length, the feature sub-vectors of the hot goods, the normal goods, and the cold goods may also be set to be a preset dimension (e.g., 10, 8, 5) with different lengths, respectively, and so on. The preset dimension may be set according to the data characteristics of the second class of objects, for example, in table 1, the first class of commodities are regarded as cold commodities, the second class of commodities are regarded as hot commodities, and the third class of commodities are regarded as normal commodities. The preset dimensionality of hot commodities can be set to be 5 (browsing times, purchase adding times, comment times, order placing times and attention degrees respectively) according to manual setting, the preset dimensionality of normal commodities is 4 (browsing times, purchase adding times, order placing times and attention degrees respectively), the preset dimensionality of cold commodities is 4 (browsing times, comment times, order placing times and attention degrees respectively), and further, each group of feature data can be converted into a feature sub-vector according to the preset dimensionality. For example, according to the setting of the preset dimensions of the hot goods, the normal goods and the cold goods in the above example, the feature data of the user a about the hot goods may be converted into the feature sub-vectors [80, 80, 13, 25, 25], the feature data about the normal goods may be converted into the feature sub-vectors [20, 10, 5, 20], the feature data about the cold goods may be converted into the feature sub-vectors [10, 5, 2, 3], and the feature sub-vectors may be spliced to obtain the feature vectors [80, 80, 13, 25, 25, 20, 10, 5, 20, 10, 5, 2, 3 ].
It should be noted that, in the exemplary embodiment, a preset dimension may also be determined according to the characteristics of the acquired feature data of the first-class object, for example, the browsing amount of the hot commodity is far greater than that of the normal commodity and the cold commodity, and the browsing times of the hot commodity by the user are relatively small, so that it is difficult to perform personalized distinction on the user through the data of the dimension of the browsing times, and therefore, whether to perform feature sub-vector conversion on the data of the dimension of the browsing times and perform subsequent calculation has little influence on the object recommendation method in the exemplary embodiment. Similarly, the preset dimensions of the cold commodity and the normal commodity may also be set according to the concrete representation of the user feature data, which is not particularly limited in the present disclosure.
In an exemplary embodiment, when the preset dimension of each group of feature data is obtained, and each group of feature data is converted into the feature sub-vectors according to the preset dimension, the feature sub-vectors may be determined according to an alternating least square method.
The alternating least square method is a common calculation method in a collaborative filtering algorithm, and can be generally used for recommending similar objects, for example, recommending suitable commodities according to the preference of a user. In the present exemplary embodiment, evaluation matrices of the first class object and the second class object under multiple categories may be constructed according to multiple categories of the second class object, for example, a user-commodity matrix of a hot commodity, a user-commodity matrix of a normal commodity, a user-commodity matrix of a cold commodity, and the like may be respectively constructed. According to the least square method, each type of evaluation matrix is decomposed into a matrix related to a first type of object and a matrix related to a second type of object, and multi-dimensional characteristic data of the first type of object is obtained, for example, a user-commodity matrix of popular commodities is decomposed into a user matrix and a commodity matrix, and characteristic data related to the browsing times, the purchase adding times or the comment times is extracted from the user matrix, or characteristic data related to the browsing amount and the sales amount is extracted from the commodity matrix, and the like. According to the extracted feature data, conversion to the feature sub-vectors can be completed according to preset dimensionality.
In step S140, similarity between feature vectors of a plurality of first-class objects is determined.
And step S150, determining the first class of objects for recommendation according to the similarity.
In the present exemplary embodiment, the similarity calculation may be achieved by calculating the distance or angle between the feature vectors. The similarity can be determined by various methods, for example, the similarity of the first class of objects is determined by calculating the euclidean distance, the manhattan distance, the minkowski distance or the cosine similarity, and the smaller the distance or the angle, the higher the similarity is. Taking the euclidean distance as an example, the eigenvectors of the first class of objects may be projected in a preset space, the euclidean distance values between the projected eigenvectors are calculated, and all the calculated distance values are sorted from large to small, with a smaller distance value indicating a higher similarity between the two first class of objects. In the present exemplary embodiment, the similarity between two feature vectors may be calculated, the similarity between one feature vector and a plurality of feature vectors may be calculated, and the like.
In this exemplary embodiment, a preset threshold may be set for the similarity, and according to the calculation result of the similarity obtained in step S140, the feature vectors exceeding the threshold are clustered to obtain one or more sets of feature vectors, and the feature vector in each set may be regarded as the first class object corresponding to the feature vector with the greater similarity. The first class objects used for recommendation may be determined from the set as needed, for example, the number of the first class objects to be recommended may be set to 50, then 50 first class objects may be randomly recommended from one similarity set, or the similarities of all feature vectors in the similarity set may be ranked, the first class objects with the similarity ranked at the top 50 are recommended, and so on.
In an exemplary embodiment, step S140 may include the steps of:
step S210, obtaining a reference object from a plurality of first-class objects;
step S220, determining the similarity between the feature vectors of each first-class object and the reference object;
further, step S150 may include:
in step S230, the first class object whose similarity with the feature vector of the reference object reaches the second threshold is determined as the first class object for recommendation.
The reference object is a reference object used for determining the recommended object, for example, if a business wishes to determine other users that can be extended through behavior data of existing users, the existing users may be regarded as the reference object. The reference object is usually contained in a plurality of objects of the first type, so that one or more objects can be determined from the objects of the first type as reference objects. By taking the reference object as a reference, the similarity between the reference object and other first-class objects can be respectively calculated, and the calculated similarities can be sorted. When the first-class object for recommendation is determined, a second threshold value can be set, and the first-class object meeting the second threshold value is used as the first-class object for recommendation. The second threshold may be a similarity value set artificially according to needs, and all or part of the objects of the first class higher than the similarity value may be used as the first class of objects that can be recommended, or the second threshold may be determined according to the second class of objects that needs to be recommended, for example, 50 users need to be recommended, and a value with the lowest similarity among the 50 users may be used as the second threshold, and the like.
Based on the above description, in the present exemplary embodiment, the feature data of the first-class object is obtained, and based on the association between the first-class object and the multiple classes of the second-class object, the feature data of each first-class object is divided into multiple groups of feature data, and is converted into feature sub-vectors, so as to determine the feature vectors, and then the first-class object for recommendation is determined according to the similarity between the first objects. On one hand, compared with a general recommendation method in which feature data is directly processed, the exemplary embodiment preprocesses the feature data based on the association between the feature data of the first class of objects and multiple classes of the second class of objects, determines the characteristics of different first class of objects, and makes object recommendation more targeted and effective; on the other hand, the feature data of the first-class object is divided into a plurality of groups of feature data, the feature vectors are determined after the feature sub-vectors are converted, and the feature vectors of the first-class object which are more abundant and effective can be obtained, so that accurate recommendation can be performed.
In an exemplary embodiment, step S140 may include the steps of:
step S310, clustering a plurality of first-class objects based on the similarity;
step S320, determining a first class object for recommendation according to the result of clustering.
In the exemplary embodiment, the similarity between the feature vectors can be obtained by calculating the feature vector of each first-class object, and the first-class objects corresponding to the feature vectors with the similarity reaching the preset condition can be clustered based on the similarity to obtain a clustering result. According to the clustering result, the first-class objects with the same clustering result can be recommended to each other, for example, if the clustering results of the user a, the user B and the user C are the same, the user B and the user C can be recommended to the service provider of the user a as the recommended objects, or the user a and the user B can be recommended to the service provider of the user C as the recommended objects. In addition, a first class object may be determined, and objects for recommendation may be determined according to the first class object, for example, the clustering results of the user a, the user B, and the user C are the same, where the user a is an existing user determined by the service provider, and then the user B and the user C may be used as objects for recommendation. There may be a plurality of first-class objects for recommendation, and the specific number may have a plurality of setting manners, for example, the similarity calculated in step S310 may be ranked, when the first-class object for recommendation is determined, the first-class object for recommendation is determined according to the ranking result, the first-class object whose similarity is higher than a certain threshold may also be used as the first-class object for recommendation, and the like.
Further, in an exemplary embodiment, the step S320 may include the steps of:
step S410, performing first clustering on each first class object based on the characteristic vector of each first class object to obtain a plurality of sets related to the first class objects;
step S420, respectively judging whether each set meets preset conditions;
and step S430, if the set meets the preset condition, clustering the first class objects in the set again.
In the exemplary embodiment, considering that under different clustering conditions, the first class objects in the same clustering result may appear in too many numbers, and the clustering result is unbalanced, which results in complicated calculation, a twice clustering manner may be adopted, and based on the feature vector of each first class object, a K-means clustering manner may be firstly adopted to perform first clustering on each first class object to obtain a plurality of sets of first class objects, where each set is the first class object with higher similarity. The preset condition may be a judgment criterion for judging whether the first class objects in each set need to be clustered again, the preset condition may be a preset threshold of similarity, the first class objects which do not reach the preset threshold are clustered again, the preset condition may also be the number of the first class objects in the set, if the number exceeds a certain preset threshold, the first class objects in the set may be clustered again, and so on. It should be noted that, in the exemplary embodiment, besides the K-means clustering method, there may be multiple ways, such as mean shift clustering, density-based clustering method, maximum expected clustering based on a gaussian mixture model, and the like, which is not particularly limited by the present disclosure.
In an exemplary embodiment, the preset condition is that the number of the first type objects in the set is less than a first threshold, and step S430 may include:
and if the number of the first class objects in the set reaches a first threshold value, determining a hash parameter according to the number of the first class objects in the set, and clustering the first class objects in the set again by adopting a locality sensitive hash algorithm.
The first threshold is a condition for judging whether the number of the first class objects in the set is appropriate, and if the number of the first class objects in the set reaches the first threshold, the number of the first class objects in the set is large, so that the first class objects in the set can be clustered again for convenience of subsequent calculation. The clustering method can adopt a locality sensitive hashing algorithm, hash parameters are determined according to the number of first-class objects in a set, the first-class objects in the set can be mapped into a plurality of hash buckets by adjusting the hash parameters, the hash parameters can comprise projection lengths and projection directions, the probability that the first-class objects are mapped into the same hash bucket is higher if the projection lengths are longer or the projection directions are more, and longer projection lengths or more projection directions can be set if the number of the first-class objects obtained by first clustering is more. The first class objects stored in each hash bucket comprise nearest neighbor first class objects, and if a first class object is given, the first class object can be mapped into a certain hash bucket through a locality sensitive hashing algorithm, the first class objects stored in the hash bucket form the nearest neighbor first class objects which can be used for recommendation, and therefore accurate recommendation of the first class objects can be completed.
In an exemplary embodiment, the association relationship between the feature data of the first class object and the multiple classes of the second class object may be obtained by obtaining statistics of the feature data of part or all of the first class objects on the platform, or may be obtained directly according to the analysis of the feature data of the first class objects obtained in step S110. In the present exemplary embodiment, the object recommendation method may further include:
counting the data of the second class of objects according to the characteristic data of the plurality of first class of objects;
a plurality of classes of the second class object are determined based on the data of the second class object.
For example, the feature data of the user may be obtained in step S110, the feature data about the product may be counted according to the feature data, for example, the feature data of the user a shown in table 1 is used, if the product with the browsing amount of 30 is a hot product, the second type of product is a hot product, the product with the browsing amount of less than 15 is a cold product, the first type of product is a cold product, and the third type of product is a normal product, the counting of the product data according to the feature data of the user is completed according to the above analysis, and the category of the product is determined according to the result of the counting. The category of the second class object can also be determined according to other indexes, such as purchase times, comment times, order placing times and the like. After determining which product is a hot product, in step S120, the feature data of each user may be divided into a plurality of sets of feature data based on the association between the feature data and the plurality of categories of the product.
Fig. 5 shows a flowchart of an object recommendation method in the present exemplary embodiment, including the following steps:
step S510, acquiring characteristic data of a plurality of first-class objects;
step S520, grouping the characteristic data according to the association between the characteristic data and the category of the second class object;
step S530, converting each group of feature data into feature sub-vectors, and determining the feature vectors according to the plurality of feature sub-vectors;
step S540, according to the characteristic data of the first class object obtained in the step S530, performing first clustering on the first class object to obtain a set of a plurality of characteristic vectors;
step S550, judging whether each set meets preset conditions;
step S560, if the set meets the preset condition, re-clustering the first class objects corresponding to the feature vectors in the set;
step S570, obtaining similarity ranking of the first type of objects;
step S580, determining a reference object;
in step S590, the first class object for recommendation is determined according to the similarity ranking in step S570 and the reference object in step S580.
Wherein, step S510 exemplarily shows that the feature data of the first class object a, the second class object B, and the third class object C are obtained, it should be understood that the present exemplary embodiment is not limited to obtaining only the feature data of the three first class objects, and the number of the specifically obtained first class objects is not particularly limited in this disclosure. Step S530 may determine feature data of a first type of object by using an alternating least square in a collaborative filtering algorithm and perform conversion of a feature sub-vector, which may be regarded as a process of vectorizing the first type of object, step S530 exemplarily shows a specific determination process of a feature vector of the first type of object, and the determination methods of feature vectors of other first type of objects are similar thereto, and a plurality of feature vectors corresponding to the first type of object may be obtained through step S530, as shown in the figure. In step S540, a K-means clustering method may be used to perform first clustering on the first class objects corresponding to the obtained feature vectors to obtain a plurality of sets including the feature vectors of the first class objects, in step S550, if the sets do not meet preset conditions, the step S570 may be directly performed to obtain similarity ranking of the first class objects, the step S560 is not performed to perform re-clustering on the first class objects, and a partial sensitive hashing algorithm may be used in the process of re-clustering, specifically, a bucket partitioning algorithm may be used, and the process of re-clustering is completed by mapping the adjacent first class objects into the same hash bucket. In the exemplary embodiment, data processing is performed by adopting collaborative filtering, K-means clustering and a locality sensitive hashing algorithm at each stage according to data characteristics, so that the problems of complex calculation process and memory overflow caused by large data volume and unbalanced distribution are solved, and the object recommendation efficiency is improved.
Exemplary embodiments of the present disclosure also provide an object recommending apparatus. Referring to fig. 6, the apparatus 600 may include a data acquisition module 610, a data grouping module 620, a vector determination module 630, a similarity determination module 640, and an object determination module 650. The data acquiring module 610 is configured to acquire feature data of a plurality of first-class objects; a data grouping module 620, configured to divide the feature data of each first class object into multiple groups of feature data based on associations between the feature data and multiple classes of second class objects; a vector determining module 630, configured to convert each group of feature data of each first class object into feature sub-vectors, and determine a feature vector of each first class object according to a plurality of feature sub-vectors of each first class object; a similarity determining module 640, configured to determine similarities between feature vectors of a plurality of first class objects; and an object determination module 650, configured to determine a first class of objects for recommendation according to the similarity.
In an exemplary embodiment, the second class of objects may include N categories; the data grouping module may be configured to, for each feature data, if the feature data is associated with an ith category of the N categories, divide the feature data into ith groups of feature data to divide the feature data of each first-category object into multiple groups of feature data; and the ith group of characteristic data has a corresponding relation with the ith category of the second class object.
In an exemplary embodiment, the vector determination module may include: the dimension acquiring unit is used for acquiring the preset dimension of each group of feature data and converting each group of feature data into feature sub-vectors according to the preset dimension; and the vector splicing unit is used for splicing the characteristic sub-vectors of each first class object to obtain the characteristic vector of each first class object.
In an exemplary embodiment, the object determination module may include: the clustering unit is used for clustering the first class objects based on the similarity; and the determining unit is used for determining the first class of objects for recommendation according to the clustering result.
In an exemplary embodiment, the clustering unit may include: the pre-clustering subunit is used for performing primary clustering on each first class object based on the feature vector of each first class object to obtain a plurality of sets related to the first class objects; the judging subunit is used for respectively judging whether each set meets a preset condition; and the re-clustering subunit is used for re-clustering the first class of objects in the set if the set meets the preset condition.
In an exemplary embodiment, the preset condition is that the number of the first type objects in the set is less than a first threshold; the re-clustering subunit may be configured to, if the number of the first class objects in the set reaches a first threshold, determine a hash parameter according to the number of the first class objects in the set, and perform re-clustering on the first class objects in the set by using a locality sensitive hash algorithm.
In an exemplary embodiment, the similarity determination module may include: a reference object acquisition unit for acquiring a reference object among a plurality of objects of a first type; the similarity determining unit is used for determining the similarity between the feature vectors of each first class object and the reference object; the object determination module may include: and the threshold judging unit is used for determining the first class of objects with the similarity reaching a second threshold with the feature vector of the reference object as the first class of objects for recommendation.
The specific details of each module/unit have been described in detail in the corresponding method embodiment, and therefore are not described herein again.
Exemplary embodiments of the present disclosure also provide an electronic device capable of implementing the above method.
As will be appreciated by one skilled in the art, aspects of the present disclosure may be embodied as a system, method or program product. Accordingly, various aspects of the present disclosure may be embodied in the form of: an entirely hardware embodiment, an entirely software embodiment (including firmware, microcode, etc.) or an embodiment combining hardware and software aspects that may all generally be referred to herein as a "circuit," module "or" system.
An electronic device 700 according to such an exemplary embodiment of the present disclosure is described below with reference to fig. 7. The electronic device 700 shown in fig. 7 is only an example and should not bring any limitation to the functions and the scope of use of the embodiments of the present disclosure.
As shown in fig. 7, electronic device 700 is embodied in the form of a general purpose computing device. The components of the electronic device 700 may include, but are not limited to: the at least one processing unit 710, the at least one memory unit 720, a bus 730 connecting different system components (including the memory unit 720 and the processing unit 710), and a display unit 740.
Where the memory unit stores program code, the program code may be executed by the processing unit 710 such that the processing unit 710 performs the steps according to various exemplary embodiments of the present disclosure as described in the above-mentioned "exemplary methods" section of this specification. For example, the processing unit 710 may execute steps S110 to S150 shown in fig. 1, or may execute steps S310 to S320 shown in fig. 3, and the like.
The storage unit 720 may include readable media in the form of volatile memory units, such as a random access memory unit (RAM)721 and/or a cache memory unit 722, and may further include a read only memory unit (ROM) 723.
The memory unit 720 may also include programs/utilities 724 having a set (at least one) of program modules 725, such program modules 725 including, but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which, or some combination thereof, may comprise an implementation of a network environment.
Bus 730 may be any representation of one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.
The electronic device 700 may also communicate with one or more external devices 900 (e.g., keyboard, pointing device, bluetooth device, etc.), with one or more devices that enable a user to interact with the electronic device 700, and/or with any devices (e.g., router, modem, etc.) that enable the electronic device 700 to communicate with one or more other computing devices. Such communication may occur via an input/output (I/O) interface 750. Also, the electronic device 700 may communicate with one or more networks (e.g., a Local Area Network (LAN), a Wide Area Network (WAN), and/or a public network such as the internet) via the network adapter 760. As shown, the network adapter 760 communicates with the other modules of the electronic device 700 via the bus 730. It should be appreciated that although not shown in the figures, other hardware and/or software modules may be used in conjunction with the electronic device 700, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, among others.
Through the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein may be implemented by software, or by software in combination with necessary hardware. Therefore, the technical solution according to the embodiments of the present disclosure may be embodied in the form of a software product, which may be stored in a non-volatile storage medium (which may be a CD-ROM, a usb disk, a removable hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which may be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to the exemplary embodiments of the present disclosure.
Exemplary embodiments of the present disclosure also provide a computer-readable storage medium having stored thereon a program product capable of implementing the above-described method of the present specification. In some possible embodiments, various aspects of the disclosure may also be implemented in the form of a program product comprising program code for causing a terminal device to perform the steps according to various exemplary embodiments of the disclosure described in the above-mentioned "exemplary methods" section of this specification, when the program product is run on the terminal device.
Referring to fig. 8, a program product 800 for implementing the above method according to an exemplary embodiment of the present disclosure is described, which may employ a portable compact disc read only memory (CD-ROM) and include program code, and may be run on a terminal device, such as a personal computer. However, the program product of the present disclosure is not limited thereto, and in this document, a 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.
The program product may employ any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. A 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 (a non-exhaustive list) of the readable storage medium include: an electrical connection having one or more wires, a portable disk, 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.
A computer readable signal medium may include a propagated data signal with 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 readable signal medium may also be any readable medium that is not a 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 readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.
Program code for carrying out operations for the present disclosure may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's computing device and partly on a remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computing device (e.g., through the internet using an internet service provider).
Furthermore, the above-described figures are merely schematic illustrations of processes included in methods according to exemplary embodiments of the present disclosure, and are not intended to be limiting. It will be readily understood that the processes shown in the above figures are not intended to indicate or limit the chronological order of the processes. In addition, it is also readily understood that these processes may be performed synchronously or asynchronously, e.g., in multiple modules.
It should be noted that although in the above detailed description several modules or units of the device for action execution are mentioned, such a division is not mandatory. Indeed, the features and functions of two or more modules or units described above may be embodied in one module or unit according to an exemplary embodiment of the present disclosure. Conversely, the features and functions of one module or unit described above may be further divided into embodiments by a plurality of modules or units.
Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure disclosed herein. This application is intended to cover any variations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.
It will be understood that the present disclosure is not limited to the precise arrangements described above and shown in the drawings and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is to be limited only by the terms of the appended claims.

Claims (10)

1. An object recommendation method, comprising:
acquiring characteristic data of a plurality of first-class objects;
dividing the feature data of each first class object into a plurality of groups of feature data based on the association between the feature data and a plurality of classes of second class objects;
converting each group of feature data of each first class object into a feature sub-vector respectively, and determining the feature vector of each first class object according to a plurality of feature sub-vectors of each first class object;
determining similarity between feature vectors of the plurality of first class objects;
and determining a first class of objects for recommendation according to the similarity.
2. The method of claim 1, wherein the second class of objects comprises N classes;
the dividing the feature data of each first class object into a plurality of groups of feature data based on the association between the feature data and a plurality of classes of second class objects comprises:
for each feature data, if the feature data is associated with the ith category in the N categories, dividing the feature data into ith groups of feature data so as to divide the feature data of each first-class object into multiple groups of feature data;
and the ith group of characteristic data has a corresponding relation with the ith category of the second class object.
3. The method according to claim 1, wherein converting each set of feature data of each first-class object into a feature sub-vector, and determining the feature vector of each first-class object according to a plurality of feature sub-vectors of each first-class object comprises:
acquiring a preset dimension of each group of feature data, and converting each group of feature data into a feature sub-vector according to the preset dimension;
and splicing the characteristic sub-vectors of each first class object to obtain the characteristic vector of each first class object.
4. The method of claim 1, wherein the determining the first class of objects for recommendation based on the similarity comprises:
clustering the plurality of first class objects based on the similarity;
and determining the first class of objects for recommendation according to the clustering result.
5. The method of claim 4, wherein clustering the plurality of first class objects based on the similarity comprises:
clustering each first class object for the first time based on the feature vector of each first class object to obtain a plurality of sets of the first class objects;
respectively judging whether each set meets a preset condition;
and if the set meets the preset condition, clustering the first class of objects in the set again.
6. The method according to claim 5, wherein the preset condition is that the number of the first type objects in the set is less than a first threshold;
if the set meets the preset condition, clustering the first class of objects in the set again comprises:
and if the number of the first class objects in the set reaches the first threshold value, determining a hash parameter according to the number of the first class objects in the set, and clustering the first class objects in the set again by adopting a locality sensitive hash algorithm.
7. The method of claim 1, wherein determining the similarity between the feature vectors of the plurality of objects of the first class comprises:
acquiring a reference object in the plurality of first-class objects;
determining similarity between feature vectors of each first class object and the reference object;
the determining the first class of objects for recommendation according to the similarity includes:
and determining the first class of objects with the similarity reaching a second threshold value with the feature vector of the reference object as the first class of objects for recommendation.
8. An object recommendation apparatus, comprising:
the data acquisition module is used for acquiring characteristic data of a plurality of first-class objects;
the data grouping module is used for dividing the characteristic data of each first class object into a plurality of groups of characteristic data based on the association between the characteristic data and a plurality of classes of second class objects;
the vector determination module is used for converting each group of feature data of each first class object into feature sub-vectors respectively and determining the feature vector of each first class object according to the plurality of feature sub-vectors of each first class object;
the similarity determining module is used for determining the similarity among the feature vectors of the plurality of first-class objects;
and the object determining module is used for determining the first class of objects for recommendation according to the similarity.
9. An electronic device, comprising:
a processor; and
a memory for storing executable instructions of the processor;
wherein the processor is configured to perform the method of any of claims 1-7 via execution of the executable instructions.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the method of any one of claims 1 to 7.
CN201910420177.XA 2019-05-20 2019-05-20 Object recommendation method, object recommendation device, electronic equipment and storage medium Pending CN111966886A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910420177.XA CN111966886A (en) 2019-05-20 2019-05-20 Object recommendation method, object recommendation device, electronic equipment and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910420177.XA CN111966886A (en) 2019-05-20 2019-05-20 Object recommendation method, object recommendation device, electronic equipment and storage medium

Publications (1)

Publication Number Publication Date
CN111966886A true CN111966886A (en) 2020-11-20

Family

ID=73357934

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910420177.XA Pending CN111966886A (en) 2019-05-20 2019-05-20 Object recommendation method, object recommendation device, electronic equipment and storage medium

Country Status (1)

Country Link
CN (1) CN111966886A (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112348123A (en) * 2020-12-08 2021-02-09 武汉卓尔数字传媒科技有限公司 User clustering method and device and electronic equipment
CN112925993A (en) * 2021-04-08 2021-06-08 国网电子商务有限公司 Collaborative filtering recommendation method and device, storage medium and equipment
CN113793180A (en) * 2021-09-15 2021-12-14 北京沃东天骏信息技术有限公司 User preference analysis method, device, equipment and computer storage medium
CN114547482A (en) * 2022-03-03 2022-05-27 智慧足迹数据科技有限公司 Service feature generation method and device, electronic equipment and storage medium
CN116342267A (en) * 2022-12-29 2023-06-27 中央国债登记结算有限责任公司深圳分公司 Method and device for analyzing similarity of underwriter behaviors, storage medium and electronic equipment
CN116342229A (en) * 2023-05-30 2023-06-27 日照金果粮油有限公司 Cross-border electronic commerce information transaction processing system

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112348123A (en) * 2020-12-08 2021-02-09 武汉卓尔数字传媒科技有限公司 User clustering method and device and electronic equipment
CN112925993A (en) * 2021-04-08 2021-06-08 国网电子商务有限公司 Collaborative filtering recommendation method and device, storage medium and equipment
CN113793180A (en) * 2021-09-15 2021-12-14 北京沃东天骏信息技术有限公司 User preference analysis method, device, equipment and computer storage medium
CN114547482A (en) * 2022-03-03 2022-05-27 智慧足迹数据科技有限公司 Service feature generation method and device, electronic equipment and storage medium
CN114547482B (en) * 2022-03-03 2023-01-20 智慧足迹数据科技有限公司 Service feature generation method and device, electronic equipment and storage medium
CN116342267A (en) * 2022-12-29 2023-06-27 中央国债登记结算有限责任公司深圳分公司 Method and device for analyzing similarity of underwriter behaviors, storage medium and electronic equipment
CN116342229A (en) * 2023-05-30 2023-06-27 日照金果粮油有限公司 Cross-border electronic commerce information transaction processing system
CN116342229B (en) * 2023-05-30 2023-08-08 日照金果粮油有限公司 Cross-border electronic commerce information transaction processing system

Similar Documents

Publication Publication Date Title
US10410138B2 (en) System and method for automatic generation of features from datasets for use in an automated machine learning process
CN110866181B (en) Resource recommendation method, device and storage medium
US11682093B2 (en) Document term recognition and analytics
CN111966886A (en) Object recommendation method, object recommendation device, electronic equipment and storage medium
Thorleuchter et al. Analyzing existing customers’ websites to improve the customer acquisition process as well as the profitability prediction in B-to-B marketing
US9911131B1 (en) Method and system for obtaining leads based on data derived from a variety of sources
US20170091847A1 (en) Automated feature identification based on review mapping
WO2021174944A1 (en) Message push method based on target activity, and related device
KR20160121806A (en) Determining a temporary transaction limit
EP4322031A1 (en) Recommendation method, recommendation model training method, and related product
US20190080352A1 (en) Segment Extension Based on Lookalike Selection
CN107908616B (en) Method and device for predicting trend words
CN111612581A (en) Method, device and equipment for recommending articles and storage medium
CN114997916A (en) Prediction method, system, electronic device and storage medium of potential user
US11334935B2 (en) Method, system, and manufacture for light hypergraph based recommendation
CN113886697A (en) Clustering algorithm based activity recommendation method, device, equipment and storage medium
KR102299525B1 (en) Product Evolution Mining Method And Apparatus Thereof
US20230245210A1 (en) Knowledge graph-based information recommendation
CN113837843B (en) Product recommendation method and device, medium and electronic equipment
US20140324523A1 (en) Missing String Compensation In Capped Customer Linkage Model
US20140324524A1 (en) Evolving a capped customer linkage model using genetic models
US11487964B2 (en) Comprehensive data science solution for segmentation analysis
CN113792952A (en) Method and apparatus for generating a model
CN114429384A (en) Intelligent product recommendation method and system based on e-commerce platform
CN112967100A (en) Similar population expansion method, device, computing equipment and medium

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination