CN109978580A - Object recommendation method, apparatus and computer readable storage medium - Google Patents

Object recommendation method, apparatus and computer readable storage medium Download PDF

Info

Publication number
CN109978580A
CN109978580A CN201711455376.1A CN201711455376A CN109978580A CN 109978580 A CN109978580 A CN 109978580A CN 201711455376 A CN201711455376 A CN 201711455376A CN 109978580 A CN109978580 A CN 109978580A
Authority
CN
China
Prior art keywords
attribute
click
recommendation
browsing
determining
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
CN201711455376.1A
Other languages
Chinese (zh)
Inventor
郭雪茹
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 CN201711455376.1A priority Critical patent/CN109978580A/en
Publication of CN109978580A publication Critical patent/CN109978580A/en
Pending legal-status Critical Current

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION 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
    • G06Q30/00Commerce
    • G06Q30/02Marketing; Price estimation or determination; Fundraising
    • G06Q30/0241Advertisements
    • G06Q30/0251Targeted advertisements
    • G06Q30/0255Targeted advertisements based on user history
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION 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
    • G06Q30/00Commerce
    • G06Q30/02Marketing; Price estimation or determination; Fundraising
    • G06Q30/0241Advertisements
    • G06Q30/0251Targeted advertisements
    • G06Q30/0269Targeted advertisements based on user profile or attribute
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION 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
    • G06Q30/00Commerce
    • G06Q30/06Buying, selling or leasing transactions
    • G06Q30/0601Electronic shopping [e-shopping]
    • G06Q30/0631Item recommendations

Landscapes

  • Business, Economics & Management (AREA)
  • Accounting & Taxation (AREA)
  • Finance (AREA)
  • Strategic Management (AREA)
  • Engineering & Computer Science (AREA)
  • Development Economics (AREA)
  • General Physics & Mathematics (AREA)
  • Economics (AREA)
  • Marketing (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • Game Theory and Decision Science (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

This disclosure relates to which a kind of object recommendation method, apparatus and computer readable storage medium, are related to field of computer technology.Disclosed method includes: to obtain the reference object combination with the reference object composition for recommending relevance;According to the similarity of each attribute value of the significance level of each attribute and alternative objects attribute value corresponding with reference object, determine alternative objects similar with reference object as recommended;According to reference object combine in the corresponding relationship of reference object determine the corresponding relationship of recommended, obtain recommended combination, recommend to be combined into user according to recommended group.The scheme of the disclosure can refer to the use habit and interest of user, be more accurately completely user's recommended, and solve the problems, such as that long-tail commodity are not recommended to a certain extent, promote user experience.

Description

Object recommendation method, device and computer-readable storage medium
Technical Field
The present disclosure relates to the field of computer technologies, and in particular, to an object recommendation method and apparatus, and a computer-readable storage medium.
Background
With the development of internet technology, people have become more and more popular to perform daily activities such as entertainment, social contact, shopping and the like through networks.
It has become a relatively common phenomenon that internet platforms attract users through smart marketing. The internet platform will generally recommend products for the user according to the purchasing or using habits of the user, for example, recommend related goods, games, news, videos or microblogs for the user.
The internet platform can launch various new special services along with the development of services. For these feature services, the data volume has not yet reached a certain level, and the recommendation directly based on the data of the feature services may be inaccurate. Currently, popular products are generally recommended directly for users.
Disclosure of Invention
The inventor finds that: the method has the advantages that popular products are directly recommended for users, the interests of the users are not considered, the recommendation result is inaccurate and incomplete, the recommendation of long-tail commodities is neglected, the users are easy to fatigue and lose interests, and the user experience is poor.
One technical problem to be solved by the present disclosure is: the product is recommended to the user more accurately and comprehensively when the data volume is small, and the user experience is improved.
According to some embodiments of the present disclosure, there is provided an object recommendation method including: acquiring a reference object combination consisting of reference objects with recommendation relevance; determining the alternative object similar to the reference object as a recommendation object according to the importance degree of each attribute and the similarity of each attribute value of the alternative object and the attribute value corresponding to the reference object; and determining the corresponding relation of the recommended objects according to the corresponding relation of the reference objects in the reference object combination to obtain the recommended object combination so as to recommend the user according to the recommended object combination.
In some embodiments, the importance of the attributes is determined using the following method: determining the click rate of the attribute as the importance degree of the attribute according to the click times and the number of click users of each label corresponding to the attribute; or determining the browsing dispersion degree of the attribute as the importance degree of the attribute according to the number of the labels with the same attribute and the object browsing amount corresponding to each label.
In some embodiments, the importance of the attributes is determined using the following method: determining the click rate of the attribute according to the click times and the number of click users of each label corresponding to the attribute; determining the browsing dispersion degree of the attribute according to the number of the tags of the attribute and the browsing amount of the object corresponding to each tag; and determining the importance degree of the attribute according to the click rate and the browsing dispersion degree.
In some embodiments, the click rate of the attribute is a weighted sum of a total number of clicks of each tag corresponding to the attribute and a total number of users clicked.
In some embodiments, the number of clicks and the number of click users are the number of clicks and the number of click users of valid clicks; the effective click is determined according to click lead-in information, wherein the click lead-in information comprises the following components: at least one of the number of objects browsed by click-through, the average browsing volume per object by click-through, and the average browsing duration per object by click-through.
In some embodiments, the browsing dispersion degree of the attribute is a standard deviation of the browsing amount of the object corresponding to each tag of the attribute.
In some embodiments, the candidate objects similar to the reference object are determined using the following method: respectively multiplying the similarity of each attribute value of the candidate object and each attribute value of the reference object by the importance degree of the corresponding attribute to combine the similarity into an attribute vector of the candidate object; combining the importance degrees of the attributes into an attribute vector of a reference object; and determining the candidate objects similar to the reference object according to the similarity of the attribute vectors of the candidate objects and the attribute vector of the reference object.
In some embodiments, the reference object combination is obtained by performing relevance analysis on each object according to a frequent pattern growth FP-grow algorithm.
According to other embodiments of the present disclosure, there is provided an object recommending apparatus including: the reference object determining module is used for acquiring a reference object combination consisting of reference objects with recommendation relevance; the recommendation object determining module is used for determining the alternative object similar to the reference object as the recommendation object according to the importance degree of each attribute and the similarity between each attribute value of the alternative object and the corresponding attribute value of the reference object; and the recommendation combination determining module is used for determining the corresponding relation of the recommendation objects according to the corresponding relation of the reference objects in the reference object combination to obtain a recommendation object combination so as to recommend the user according to the recommendation object combination.
In some embodiments, the recommendation object determining module is configured to determine, according to the number of clicks of each tag corresponding to the attribute and the number of clicks, a click rate of the attribute as an importance degree of the attribute; or determining the browsing dispersion degree of the attribute as the importance degree of the attribute according to the number of the labels with the same attribute and the object browsing amount corresponding to each label.
In some embodiments, the recommended object determining module is configured to determine a click rate of the attribute according to the number of clicks of each tag corresponding to the attribute and the number of clicks, determine a browsing dispersion degree of the attribute according to the number of tags of the attribute and an object browsing amount corresponding to each tag, and determine an importance degree of the attribute according to the click rate and the browsing dispersion degree.
In some embodiments, the click rate of the attribute is a weighted sum of a total number of clicks of each tag corresponding to the attribute and a total number of users clicked.
In some embodiments, the number of clicks and the number of click users are the number of clicks and the number of click users of valid clicks; the effective click is determined according to click lead-in information, wherein the click lead-in information comprises the following components: at least one of the number of objects browsed by click-through, the average browsing volume per object by click-through, and the average browsing duration per object by click-through.
In some embodiments, the browsing dispersion degree of the attribute is a standard deviation of the browsing amount of the object corresponding to each tag of the attribute.
In some embodiments, the recommended object determining module is configured to multiply the similarity between each attribute value of the candidate object and each attribute value of the reference object by the importance degree of the corresponding attribute, and combine the similarity into an attribute vector of the candidate object; combining the importance degrees of the attributes into an attribute vector of a reference object; and determining the candidate objects similar to the reference object according to the similarity of the attribute vectors of the candidate objects and the attribute vector of the reference object.
In some embodiments, the reference object determination module is configured to perform relevance analysis on each object according to a frequent pattern growth FP-grow algorithm to obtain a reference object combination.
According to still other embodiments of the present disclosure, there is provided an object recommendation apparatus including: a memory; and a processor coupled to the memory, the processor configured to execute the object recommendation method of any of the preceding embodiments based on instructions stored in the memory device.
According to still further embodiments of the present disclosure, there is provided an object recommendation apparatus, a computer readable storage medium, having a computer program stored thereon, wherein the program, when executed by a processor, implements the steps of the object recommendation method of any of the preceding embodiments.
The method comprises the steps of preferentially acquiring a reference object combination, including reference objects with recommendation relevance, further selecting candidate objects similar to the reference objects from the candidate objects as recommendation objects, and determining the similarity degree of the two objects by referring to the importance degree of each attribute to a user while considering the similarity of the attribute values. The scheme disclosed by the invention is applied to a characteristic business scene, an object with recommendation relevance in an internet platform can be used as a reference object, an object similar to the reference object in the characteristic business is further determined to be used as a recommendation object, the object can be more accurately and comprehensively recommended for a user by referring to the use habits and interests of the user, the problem that long-tailed commodities are not recommended is solved to a certain extent, and the user experience is improved.
Other features of the present disclosure and advantages thereof will become apparent from the following detailed description of exemplary embodiments thereof, which proceeds with reference to the accompanying drawings.
Drawings
In order to more clearly illustrate the embodiments of the present disclosure or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present disclosure, and other drawings can be obtained by those skilled in the art without creative efforts.
Fig. 1 illustrates a flow diagram of an object recommendation method of some embodiments of the present disclosure.
FIG. 2 shows a flow diagram of object recommendation methods of further embodiments of the present disclosure.
Fig. 3 illustrates a schematic structural diagram of an object recommendation device according to some embodiments of the present disclosure.
Fig. 4 is a schematic structural diagram of an object recommendation apparatus according to another embodiment of the present disclosure.
Fig. 5 is a schematic structural diagram of an object recommendation apparatus according to still other embodiments of the present disclosure.
Detailed Description
The technical solutions in the embodiments of the present disclosure will be clearly and completely described below with reference to the drawings in the embodiments of the present disclosure, and it is obvious that the described embodiments are only a part of the embodiments of the present disclosure, and not all of the embodiments. The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the disclosure, its application, or uses. All other embodiments, which can be derived by a person skilled in the art from the embodiments disclosed herein without making any creative effort, shall fall within the protection scope of the present disclosure.
The scheme is provided for solving the problems that popular products are directly recommended to users in special services, the interests of the users are not considered, the recommendation results are inaccurate and incomplete, the recommendation of long-tail commodities is neglected, and the user experience is not good. The object recommendation method of the present disclosure is described below in conjunction with fig. 1.
FIG. 1 is a flow chart of some embodiments of an object recommendation method of the present disclosure. As shown in fig. 1, the method of this embodiment includes: steps S102 to S106.
Step S102, a reference object combination of reference object compositions with recommendation relevance is obtained.
The objects are, for example, various objects related to user behaviors, such as commodities, videos, news, web pages, and the like, which can be recommended to the user. The reference object with recommendation relevance may be calculated by using any existing algorithm according to a behavior record (for example, a purchase record, a browsing record, and the like within a preset number of days before and after) of the user, and for example, the reference object with recommendation relevance is obtained by performing relevance analysis on each object according to an FP-grow (frequent pattern growth) algorithm. The reference object with recommendation relevance may be obtained from data of an internet platform that develops a featured service or from data of other related platforms. For example, for an e-commerce platform, a feature service such as global purchasing may be developed, and a product with recommendation relevance may be obtained according to purchase data of the e-commerce platform.
And step S104, determining the candidate objects similar to the reference object as recommendation objects according to the importance degree of each attribute and the similarity between each attribute value of the candidate objects and the attribute value corresponding to the reference object.
When determining a candidate similar to the reference object, two factors need to be considered: on one hand, the importance degree of each attribute, namely the attributes of which aspects are more important when the user selects the object; another aspect is the similarity of each attribute value of the candidate object to the attribute value corresponding to the reference object. The candidate object may be the same or similar object as the reference object class. The alternate objects may be objects in a featured service.
The attributes of an object are used to represent the characteristics of the object. For example, for a good in an e-commerce platform, attributes include: brand, price, etc. Different attributes can be provided for different commodities, for example, for a commodity of the class of mobile phones, the attribute can be brand, price, color, screen size and the like, and for skin care products, the attribute can be brand, price, skin suitability, efficacy and the like. The internet platform may set an attribute tag for the attribute of the object (for example, on a list page, a search page, and the like), and a user may filter out the corresponding object by clicking the attribute tag, for example, the tag is price 100-. The attribute value is a specific value corresponding to the attribute, for example, the attribute value corresponding to the price attribute is 100, and the attribute value corresponding to the color is red.
The attribute value is a value of a specific object under the attribute, for example, an attribute value of a certain mobile phone under the price attribute is 3000 yuan. The method for judging the similarity of two attribute values of the same attribute can be adjusted according to actual requirements. For example, if the attribute values are the same, the similarity is 1, and if they are not the same, the similarity is 0. For some special attribute values with actual values, such as prices, the ratio of the attribute value of the candidate object to the attribute value of the reference object may be used as the similarity, for example, the ratio of the price of the candidate object to the price of the reference object is 0.8, and the similarity between the candidate object and the reference object is 0.8; or setting a difference threshold between the attribute values of the candidate objects and the reference object, setting the similarity corresponding to the candidate objects whose difference between the attribute values of the candidate objects and the reference object is within the difference threshold to be 1, and setting the similarity corresponding to the candidate objects exceeding the difference threshold to be 0. For example, the similarity corresponding to the candidate whose price gap of the reference alignment is within 20% of the price of the reference object is set to 1. Some objects may have multiple values under one attribute, for example, a contract attribute of a handset may support contracts for multiple operators, and then there may be multiple attribute values under the contract attribute. If one attribute of the reference object contains a plurality of attribute values, the attribute values of the two attributes of the candidate object can be considered to be the same as each other if one attribute value of the candidate object is the same as that of the reference object under the attribute. And are not limited to the examples given.
The importance of the attributes indicates which attributes of the object the user is more interested in when selecting the object. The objects of different classes have different attributes and the importance of each attribute is different. In some embodiments, the click rate of the attribute is determined as the importance degree of the attribute according to the click times and the number of click users of each tag corresponding to the attribute. Specifically, different weights may be set for the number of clicks and the number of clicked users, and the click degree of the attribute is a weighted sum of the total number of clicks and the total number of clicked users of each tag corresponding to the attribute.
For an object of the class, e.g., cell phone, the attributes include: labels such as Huaye, apple, Samsung and the like can be set for the attribute of the brand, labels such as 1000-2000-3000-for the attribute of the price, labels such as android, IOS and the like can be set for the attribute of the system, and labels such as 3.1-4.5 inches, 4.6-5.0 inches and the like can be set for the attribute of the screen size, so that a user can click and screen out a corresponding mobile phone. The number of clicks and the number of clicks of each label can be obtained, the total number of clicks of each label under one attribute, namely the sum of the number of clicks of each label, and the total number of clicks of each label, namely the sum of the number of clicks of each label, are determined. The degree of click of an attribute may be calculated according to the following formula.
In the formula (1), the first and second groups,for the click-through degree of the attribute i,α which is the sum of the number of clicks of each label corresponding to the attribute i1Is composed ofThe weight of (a) is determined,α which is the sum of the number of clicks of each label corresponding to the attribute i2Is composed ofI is a positive integer, and is the number of the ith attribute. Can be paired withAndnormalization is performed, for example, by dividing the index value by the maximum value of the index of this type as a normalized value,or may be a normalized value of the sum of the click times of the labels corresponding to the attribute i, that is, the attribute i is a normalized valueThe ratio of the sum of the number of clicks of each label corresponding to the attribute i to the maximum value of the sum of the number of clicks of each label corresponding to each attribute,the normalized value of the sum of the numbers of users clicking on each label corresponding to the attribute i, that is, the ratio of the sum of the numbers of users clicking on each label corresponding to the attribute i to the maximum value of the sum of the numbers of users clicking on each label corresponding to each attribute, may also be used.
Optionally, the number of clicks or the number of clicks of the tag is the number of clicks and the number of clicks of an effective click. The valid click may be determined according to click lead-in information, which includes: at least one of the number of objects browsed by click-through, the average browsing volume per object by click-through, and the average browsing duration per object by click-through.
Corresponding threshold values can be respectively set for the number of browsing objects introduced by clicking, the average browsing amount of each object introduced by clicking and the average browsing duration of each object introduced by clicking, and if one of the three items of information is lower than the corresponding threshold value, the clicking is considered to be invalid. Different judgment standards can be set according to actual conditions. And judging whether the click of the user on the label is effective click can further improve the accuracy of determining the importance degree of each attribute. Some users may click the tag to screen out the corresponding object and do not browse the object due to a click error, and the click is not included in the effective click behavior.
In some embodiments, the browsing dispersion degree of the attribute is determined as the importance degree of the attribute according to the number of the tags of the same attribute and the browsing amount of the object corresponding to each tag. Specifically, the browsing dispersion degree of the attribute is a standard deviation of the browsing amount of the object corresponding to each tag of the attribute. The browsing amount of each object can be counted, and then the label corresponding to the object is determined, so that the browsing amount of the object corresponding to each label can be further obtained. For example, as shown in table 1, for an object of the category of mobile phones, an object view volume (PV) of each tab corresponding to attributes such as a brand, a price, and a screen size can be acquired. For example, the standard deviation of the browsing amount of the object corresponding to each tag may be further calculated for the attribute of the brand.
TABLE 1
Further, the browsing dispersion degree of the attribute can be calculated according to the following formula.
In the formula (2), the first and second groups,is the browsing discrete degree of the attribute i, N is the number of labels corresponding to the attribute i, j is more than or equal to 1 and less than or equal to N, j is a positive integer,the amount of view of the tab j representing the attribute i, and μ represents the average of the amounts of view of the respective tabs of the attribute i. Can be paired withIs standardized, i.e.And dividing the sum of the browsing amount of each label corresponding to the attribute i. The higher the browsing dispersion degree of the attribute, the higher the importance degree of the attribute.
The above two embodiments for determining the importance of the attribute may be combined and supplemented, for example, if some attributes have a small number of tag clicks, but the browsing volume of the object of the attribute is large, the importance of the attribute may be measured mainly by using the browsing dispersion of the attribute. Further, the importance degree of the attribute can be determined according to the click rate and the browsing dispersion degree of the attribute. The importance of the attributes can be determined using the following formula.
In the formula (3), WiTo the degree of importance of the attribute i,for the click-through degree of the attribute i,degree of viewing dispersion for attribute i, β1Is composed ofβ weight of2 The weight of (c).
After determining the similarity of each attribute value of the candidate object to each attribute value of the reference object and the importance degree of each attribute, how to select the candidate object similar to the reference object is described below.
In some embodiments, the importance degree of the attribute is used as the weight of the corresponding attribute value, the similarity between each attribute value of the candidate object and each attribute value of the reference object is weighted to obtain the similarity between the candidate object and the reference object, and the candidate object with the highest similarity is selected as the recommended object.
In some embodiments, the similarity between each attribute value of the candidate object and each attribute value of the reference object is multiplied by the importance degree of the corresponding attribute, and the obtained similarity is combined into an attribute vector of the candidate object, and the importance degrees of the attributes are combined into an attribute vector of the reference object; and determining the candidate objects similar to the reference object according to the similarity of the attribute vectors of the candidate objects and the attribute vector of the reference object.
Referring to table 2, attributes, importance levels of the attributes, and corresponding attribute values of each object are obtained for a class object, such as a mobile phone.
TABLE 2
In this application example, with SKU (stock keeping unit) 1 as a reference object and other position candidate objects, for the attribute of price, if the price difference between SKU2 or SKU3 and SKU1 is within 20% of the price of SKU1, the similarity of the value of the price attribute of SKU2 or SKU3 and SKU1 is considered to be 1. For other attributes, if the attribute values of the two attributes are the same, the similarity is 1, otherwise, the similarity is 0. The importance levels of the respective attributes are combined into an attribute vector of (0.8, 0.6, 0.5, 0.4) of the reference object SKU 1. And respectively multiplying the similarity of each attribute value of the candidate object and each attribute value of the reference object by the importance degree of the corresponding attribute to combine the similarity into an attribute vector of the candidate object. The attribute vector for SKU2 is (0.8, 0, 0.5, 0.4) and the attribute vector for SKU3 is (0, 0.6, 0.5, 0.4). Calculating the similarity of the attribute vectors separately, the similarity of SKU2 and SKU1, and the similarity of SKU3 and SKU1 can be found. And selecting the candidate object with high similarity as a recommendation object. For example, cosine similarity between vectors is calculated as the similarity between two vectors.
For a user with long service time and more data in the internet platform, the method of the above embodiment may be adopted for the user to determine a recommended object for the user, for example, the number of clicks of each attribute by the user is determined according to the click condition of the user, the browsing amount of each label browsed by the user is determined, the importance degree of each attribute for the user is further determined, and finally, which candidate objects for the user are similar to a reference object are determined, so as to determine the recommended object.
And step S106, determining the corresponding relation of the recommended objects according to the corresponding relation of the reference objects in the reference object combination to obtain the recommended object combination so as to recommend the user according to the recommended object combination.
For example, the reference object combination is SKUA and SKUB, and if the candidate object similar to SKUA is determined to be SKUC and the candidate object similar to SKUB is determined to be SKUD, the recommended object combination is SKUC and SKUD. When the user browses or purchases the SKUC, the SKUD can be recommended to the user.
When the user is recommended, the characteristics of the user can be gathered to filter the recommended object combination. For example, if a user purchases a mobile phone within half a year, the mobile phone is not recommended for the user. And the recommendation effect can be determined according to the click rate of the recommended pit positions, and the display of the recommended object combination is sorted. The recommendation pit position is the position where the recommendation object is displayed on the page.
The method of the embodiment can also be applied to a scene that an internet platform recommends an object for a user on the basis of user data of other similar or related platforms.
In the method of the above embodiment, the reference object combination is preferentially obtained, the reference object combination includes reference objects having recommendation relevance, the candidate objects similar to the reference object are further selected from the candidate objects as recommendation objects, and the similarity degree of the two objects is determined by referring to the importance degree of each attribute to the user while considering the similarity of the attribute values. The method of the embodiment is applied to a characteristic business scene, an object with recommendation relevance in an internet platform can be used as a reference object, an object similar to the reference object in the characteristic business is further determined to be used as a recommendation object, the object can be more accurately and comprehensively recommended for a user by referring to the use habits and interests of the user, the problem that long-tailed commodities are not recommended is solved to a certain extent, and the user experience is improved.
The method of the above embodiment may be implemented by using a big data system, where the big data system includes: a big data platform, a Hadoop platform, a Spark platform, an ES (Elastic Search), a service platform, and the like. The big data platform can obtain various data from various business systems, such as user operation data, order data and the like, and stores the data for other components to call. The Hadoop platform is used for processing data in the big data platform, such as cleaning and the like. The Spark platform is used for implementing the scheme in the above embodiment to obtain the recommended object combination. The ES is used for storing the recommended object combination obtained by the Spark platform. And the service platform user calls the data in the ES to recommend the user.
Some application examples of the object recommendation method of the present disclosure are described below with reference to fig. 2.
Fig. 2 is a flow chart of some application examples of the object recommendation method of the present disclosure. As shown in fig. 2, the method of this embodiment includes: steps S202 to S214.
Step S202, determining a reference object combination composed of reference objects with recommendation relevance according to data in the Internet platform.
And step S204, selecting an object which is the same as or similar to the reference object class from the characteristic service platform as a candidate object.
Step S206, determining the importance degree of each attribute of the candidate object and the reference object, and the similarity between each attribute value of the candidate object and the corresponding attribute value of the reference object.
Step S208, determining the similarity between the candidate object and the reference object according to the importance degree of each attribute and the similarity between each attribute value of the candidate object and the corresponding attribute value of the reference object.
Step S210, selecting the object with the highest similarity to the reference object from the candidate objects as a recommendation object.
Step S212, determining the corresponding relation of the recommended objects according to the corresponding relation of the reference objects in the reference object combination to obtain the recommended object combination.
And step S214, recommending the object for the user according to the behavior of the user using the characteristic service and the recommendation rule based on the data of the recommended object combination.
The present disclosure also provides an object recommendation apparatus, which is described below in conjunction with fig. 3.
Fig. 3 is a block diagram of some embodiments of the disclosed object recommendation device. As shown in fig. 3, the apparatus 30 of this embodiment includes: a reference object determining module 302, a recommended object determining module 304, and a recommended combination determining module 306.
A reference object determining module 302, configured to obtain a reference object combination composed of reference objects with recommendation relevance. For example, the reference object determination module 302 may perform step S102 in the above-described embodiment.
In some embodiments, the reference object determining module 302 is configured to perform relevance analysis on each object according to a frequent pattern growth FP-grow algorithm to obtain a reference object combination.
And a recommended object determining module 304, configured to determine, according to the importance degree of each attribute and the similarity between each attribute value of the candidate object and the corresponding attribute value of the reference object, the candidate object similar to the reference object as the recommended object. For example, the recommended object determination module 304 may perform step S104 in the above-described embodiment.
In some embodiments, the recommended object determining module 304 is configured to multiply the similarity between each attribute value of the candidate object and each attribute value of the reference object by the importance degree of the corresponding attribute, and combine the result into an attribute vector of the candidate object; combining the importance degrees of the attributes into an attribute vector of a reference object; and determining the candidate objects similar to the reference object according to the similarity of the attribute vectors of the candidate objects and the attribute vector of the reference object.
In some embodiments, the recommendation object determining module 304 is configured to determine, according to the number of clicks of each tag corresponding to the attribute and the number of clicks, a click rate of the attribute as an importance degree of the attribute; or determining the browsing dispersion degree of the attribute as the importance degree of the attribute according to the number of the labels with the same attribute and the object browsing amount corresponding to each label.
In some embodiments, the recommended object determining module 304 is configured to determine the click rate of the attribute according to the number of clicks of each tag corresponding to the attribute and the number of clicks, determine the browsing dispersion degree of the attribute according to the number of tags of the attribute and the browsing amount of the object corresponding to each tag, and determine the importance degree of the attribute according to the click rate and the browsing dispersion degree.
Optionally, the click rate of the attribute is a weighted sum of the total click times and the total number of clicks of each tag corresponding to the attribute.
Further, the number of clicks and the number of click users are the number of clicks and the number of click users of effective clicks; the effective click is determined according to click lead-in information, wherein the click lead-in information comprises the following components: at least one of the number of objects browsed by click-through, the average browsing volume per object by click-through, and the average browsing duration per object by click-through.
Optionally, the browsing dispersion degree of the attribute is a standard deviation of browsing volumes of the objects corresponding to the tags of the attribute.
And a recommended combination determining module 306, configured to determine a corresponding relationship of the recommended objects according to the corresponding relationship of the reference objects in the reference object combinations, so as to obtain recommended object combinations, so as to recommend the users according to the recommended object combinations. For example, the recommended combination determination module 306 may perform step S106 in the above-described embodiment.
The present disclosure also provides a computer-readable storage medium having stored thereon a computer program, wherein the program, when executed by a processor, implements the steps of the object recommendation method of any of the preceding embodiments.
The object recommendation apparatus in the embodiments of the present disclosure may each be implemented by various computing devices or computer systems, which are described below in conjunction with fig. 4 and 5.
Fig. 4 is a block diagram of some embodiments of the disclosed object recommendation device. As shown in fig. 4, the apparatus 40 of this embodiment includes: a memory 410 and a processor 420 coupled to the memory 410, the processor 420 configured to perform an object recommendation method in any of the embodiments of the present disclosure based on instructions stored in the memory 410.
Memory 410 may include, for example, system memory, fixed non-volatile storage media, and the like. The system memory stores, for example, an operating system, an application program, a Boot Loader (Boot Loader), a database, and other programs.
Fig. 5 is a block diagram of another embodiment of the disclosed object recommendation device. As shown in fig. 5, the apparatus 50 of this embodiment includes: memory 510 and processor 520 are similar to memory 410 and processor 420, respectively. An input output interface 530, a network interface 540, a storage interface 550, and the like may also be included. These interfaces 530, 540, 550 and the connections between the memory 510 and the processor 520 may be, for example, via a bus 560. The input/output interface 530 provides a connection interface for input/output devices such as a display, a mouse, a keyboard, and a touch screen. The network interface 540 provides a connection interface for various networking devices, such as a database server or a cloud storage server. The storage interface 550 provides a connection interface for external storage devices such as an SD card and a usb disk.
As will be appreciated by one skilled in the art, embodiments of the present disclosure may be provided as a method, system, or computer program product. Accordingly, the present disclosure may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present disclosure may take the form of a computer program product embodied on one or more computer-usable non-transitory storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present disclosure is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the disclosure. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The above description is only exemplary of the present disclosure and is not intended to limit the present disclosure, so that any modification, equivalent replacement, or improvement made within the spirit and principle of the present disclosure should be included in the scope of the present disclosure.

Claims (18)

1. An object recommendation method comprising:
acquiring a reference object combination consisting of reference objects with recommendation relevance;
determining the alternative object similar to the reference object as a recommendation object according to the importance degree of each attribute and the similarity of each attribute value of the alternative object and the attribute value corresponding to the reference object;
and determining the corresponding relation of the recommended objects according to the corresponding relation of the reference objects in the reference object combination to obtain a recommended object combination so as to recommend the user according to the recommended object combination.
2. The object recommendation method of claim 1, wherein the importance of the attribute is determined by:
determining the click rate of the attribute as the importance degree of the attribute according to the click times and the number of click users of each label corresponding to the attribute;
or,
and determining the browsing dispersion degree of the attribute as the importance degree of the attribute according to the number of the labels with the same attribute and the object browsing amount corresponding to each label.
3. The object recommendation method of claim 1, wherein the importance of the attribute is determined by:
determining the click rate of the attribute according to the click times and the number of click users of each label corresponding to the attribute;
determining the browsing dispersion degree of the attribute according to the number of the tags of the attribute and the browsing amount of the object corresponding to each tag;
and determining the importance degree of the attribute according to the click rate and the browsing dispersion degree.
4. The object recommending method according to claim 2 or 3, wherein,
and the click rate of the attribute is the weighted sum of the total click times and the total number of click users of each label corresponding to the attribute.
5. The object recommendation method of claim 4, wherein,
the number of clicks and the number of click users are the number of clicks and the number of click users of effective clicks;
the effective click is determined according to click lead-in information, wherein the click lead-in information comprises: at least one of the number of objects browsed by click-through, the average browsing volume per object by click-through, and the average browsing duration per object by click-through.
6. The object recommending method according to claim 2 or 3, wherein,
and the browsing dispersion degree of the attribute is the standard deviation of the browsing amount of the object corresponding to each label of the attribute.
7. The object recommendation method of claim 1, wherein the candidate object similar to the reference object is determined by:
respectively multiplying the similarity of each attribute value of the candidate object and each attribute value of the reference object by the importance degree of the corresponding attribute to combine the similarity into an attribute vector of the candidate object;
combining the importance degrees of the attributes into an attribute vector of the reference object;
and determining the candidate objects similar to the reference object according to the similarity of the attribute vectors of the candidate objects and the attribute vector of the reference object.
8. The object recommendation method of claim 1, wherein,
and the reference object combination is obtained by performing relevance analysis on each object according to a frequent pattern growth FP-grow algorithm.
9. An object recommendation apparatus comprising:
the reference object determining module is used for acquiring a reference object combination consisting of reference objects with recommendation relevance;
a recommended object determining module, configured to determine, according to the importance degree of each attribute and the similarity between each attribute value of the candidate object and the attribute value corresponding to the reference object, a candidate object similar to the reference object as a recommended object;
and the recommendation combination determining module is used for determining the corresponding relation of the recommendation objects according to the corresponding relation of the reference objects in the reference object combination to obtain a recommendation object combination so as to recommend the user according to the recommendation object combination.
10. The object recommendation device of claim 9, wherein,
the recommendation object determining module is used for determining the click rate of the attribute as the importance degree of the attribute according to the click times and the number of click users of each label corresponding to the attribute; or determining the browsing dispersion degree of the attribute as the importance degree of the attribute according to the number of the labels with the same attribute and the object browsing amount corresponding to each label.
11. The object recommendation device of claim 9, wherein,
the recommended object determining module is used for determining the click rate of the attribute according to the click times and the number of click users of each label corresponding to the attribute, determining the browsing dispersion degree of the attribute according to the number of the labels of the attribute and the browsing amount of the object corresponding to each label, and determining the importance degree of the attribute according to the click rate and the browsing dispersion degree.
12. The object recommending apparatus according to claim 10 or 11, wherein,
and the click rate of the attribute is the weighted sum of the total click times and the total number of click users of each label corresponding to the attribute.
13. The object recommendation device of claim 12, wherein,
the number of clicks and the number of click users are the number of clicks and the number of click users of effective clicks;
the effective click is determined according to click lead-in information, wherein the click lead-in information comprises: at least one of the number of objects browsed by click-through, the average browsing volume per object by click-through, and the average browsing duration per object by click-through.
14. The object recommending apparatus according to claim 10 or 11, wherein,
and the browsing dispersion degree of the attribute is the standard deviation of the browsing amount of the object corresponding to each label of the attribute.
15. The object recommendation device of claim 9, wherein,
the recommended object determining module is used for multiplying the similarity of each attribute value of the candidate object and each attribute value of the reference object by the importance degree of the corresponding attribute respectively and combining the similarity into an attribute vector of the candidate object; combining the importance degrees of the attributes into an attribute vector of the reference object; and determining the candidate objects similar to the reference object according to the similarity of the attribute vectors of the candidate objects and the attribute vector of the reference object.
16. The object recommendation device of claim 9, wherein,
the reference object determining module is used for performing relevance analysis on each object according to a frequent pattern growth FP-grow algorithm to obtain a reference object combination.
17. An object recommendation apparatus comprising:
a memory; and
a processor coupled to the memory, the processor configured to perform the object recommendation method of any of claims 1-8 based on instructions stored in the memory device.
18. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 8.
CN201711455376.1A 2017-12-28 2017-12-28 Object recommendation method, apparatus and computer readable storage medium Pending CN109978580A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201711455376.1A CN109978580A (en) 2017-12-28 2017-12-28 Object recommendation method, apparatus and computer readable storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201711455376.1A CN109978580A (en) 2017-12-28 2017-12-28 Object recommendation method, apparatus and computer readable storage medium

Publications (1)

Publication Number Publication Date
CN109978580A true CN109978580A (en) 2019-07-05

Family

ID=67074156

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201711455376.1A Pending CN109978580A (en) 2017-12-28 2017-12-28 Object recommendation method, apparatus and computer readable storage medium

Country Status (1)

Country Link
CN (1) CN109978580A (en)

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110443640A (en) * 2019-07-18 2019-11-12 佛山科学技术学院 A kind of commodity method for pushing and storage medium based on big data
CN111581452A (en) * 2020-03-26 2020-08-25 浙江口碑网络技术有限公司 Method and device for obtaining recommendation object data and electronic equipment
CN111931059A (en) * 2020-08-19 2020-11-13 创新奇智(成都)科技有限公司 Object determination method and device and storage medium
CN112116421A (en) * 2020-09-09 2020-12-22 江苏亿讯网络科技有限公司 Method and system for optimizing and recommending commodities in self-built welfare mall
CN112330406A (en) * 2020-11-10 2021-02-05 深圳大学 Long-tail article recommendation method, intelligent terminal and storage medium
CN114004675A (en) * 2021-10-29 2022-02-01 北京市商汤科技开发有限公司 Object selection method and device, computer equipment and storage medium
CN114528482A (en) * 2022-01-25 2022-05-24 北京三快在线科技有限公司 Method and device for determining recommended object, electronic equipment and storage medium

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102124466A (en) * 2008-09-05 2011-07-13 索尼公司 Content recommendation system, content recommendation method, content recommendation device, program, and information storage medium
CN102542046A (en) * 2011-12-27 2012-07-04 纽海信息技术(上海)有限公司 Book recommendation method based on book contents
CN103299651A (en) * 2010-12-17 2013-09-11 北京交通大学 Program recommending device and program recommending method
CN104063481A (en) * 2014-07-02 2014-09-24 山东大学 Film individuation recommendation method based on user real-time interest vectors
CN106686460A (en) * 2016-12-22 2017-05-17 Ut斯达康(深圳)技术有限公司 Video program recommendation method and device
CN106779926A (en) * 2016-12-02 2017-05-31 乐视控股(北京)有限公司 Correlation rule generation method, device and terminal

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102124466A (en) * 2008-09-05 2011-07-13 索尼公司 Content recommendation system, content recommendation method, content recommendation device, program, and information storage medium
CN103299651A (en) * 2010-12-17 2013-09-11 北京交通大学 Program recommending device and program recommending method
CN102542046A (en) * 2011-12-27 2012-07-04 纽海信息技术(上海)有限公司 Book recommendation method based on book contents
CN104063481A (en) * 2014-07-02 2014-09-24 山东大学 Film individuation recommendation method based on user real-time interest vectors
CN106779926A (en) * 2016-12-02 2017-05-31 乐视控股(北京)有限公司 Correlation rule generation method, device and terminal
CN106686460A (en) * 2016-12-22 2017-05-17 Ut斯达康(深圳)技术有限公司 Video program recommendation method and device

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110443640A (en) * 2019-07-18 2019-11-12 佛山科学技术学院 A kind of commodity method for pushing and storage medium based on big data
CN111581452A (en) * 2020-03-26 2020-08-25 浙江口碑网络技术有限公司 Method and device for obtaining recommendation object data and electronic equipment
CN111581452B (en) * 2020-03-26 2023-10-17 浙江口碑网络技术有限公司 Recommendation object data obtaining method and device and electronic equipment
CN111931059A (en) * 2020-08-19 2020-11-13 创新奇智(成都)科技有限公司 Object determination method and device and storage medium
CN112116421A (en) * 2020-09-09 2020-12-22 江苏亿讯网络科技有限公司 Method and system for optimizing and recommending commodities in self-built welfare mall
CN112330406A (en) * 2020-11-10 2021-02-05 深圳大学 Long-tail article recommendation method, intelligent terminal and storage medium
CN112330406B (en) * 2020-11-10 2024-05-24 深圳大学 Long-tail article recommendation method, intelligent terminal and storage medium
CN114004675A (en) * 2021-10-29 2022-02-01 北京市商汤科技开发有限公司 Object selection method and device, computer equipment and storage medium
CN114528482A (en) * 2022-01-25 2022-05-24 北京三快在线科技有限公司 Method and device for determining recommended object, electronic equipment and storage medium

Similar Documents

Publication Publication Date Title
CN109978580A (en) Object recommendation method, apparatus and computer readable storage medium
CN109325179B (en) Content promotion method and device
CN108960975A (en) Personalized Precision Marketing Method, server and storage medium based on user's portrait
CN110827114A (en) Commodity recommendation method and device
CN107220852A (en) Method, device and server for determining target recommended user
WO2018121700A1 (en) Method and device for recommending application information based on installed application, terminal device, and storage medium
US10621641B2 (en) Method and device for pushing information
EP3117339A1 (en) Systems and methods for keyword suggestion
WO2018035164A1 (en) Description information generation and presentation systems, methods, and devices
CN108805598A (en) Similarity information determines method, server and computer readable storage medium
CN111626767B (en) Resource data issuing method, device and equipment
US20160196579A1 (en) Dynamic deep links based on user activity of a particular user
CN107103028A (en) A kind of information processing method and device
CN109190027A (en) Multi-source recommended method, terminal, server, computer equipment, readable medium
US9508087B1 (en) Identifying similar display items for potential placement of content items therein
CN106933905B (en) Method and device for monitoring webpage access data
CN112989276B (en) Evaluation method and device of information push system
CN106919609A (en) Product information method for pushing and device
CN110189188B (en) Commodity management method, commodity management device, computer equipment and storage medium
CN110796520A (en) Commodity recommendation method and device, computing equipment and medium
CN113158032B (en) Information pushing method and device
CN107357847B (en) Data processing method and device
CN112015970A (en) Product recommendation method, related equipment and computer storage medium
CN116308663A (en) Recommendation method, terminal and storage medium
JP5869931B2 (en) Recommendation item detection program, apparatus and method for detecting items to be criticized by critics

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
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20190705