WO2019085333A1 - 项目推荐方法、电子设备及计算机可读存储介质 - Google Patents

项目推荐方法、电子设备及计算机可读存储介质 Download PDF

Info

Publication number
WO2019085333A1
WO2019085333A1 PCT/CN2018/076171 CN2018076171W WO2019085333A1 WO 2019085333 A1 WO2019085333 A1 WO 2019085333A1 CN 2018076171 W CN2018076171 W CN 2018076171W WO 2019085333 A1 WO2019085333 A1 WO 2019085333A1
Authority
WO
WIPO (PCT)
Prior art keywords
user
specific user
type
users
specific
Prior art date
Application number
PCT/CN2018/076171
Other languages
English (en)
French (fr)
Inventor
李芳�
王建明
肖京
Original Assignee
平安科技(深圳)有限公司
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 平安科技(深圳)有限公司 filed Critical 平安科技(深圳)有限公司
Publication of WO2019085333A1 publication Critical patent/WO2019085333A1/zh

Links

Images

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
    • G06Q50/00Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/01Social networking
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90Details of database functions independent of the retrieved data types
    • G06F16/95Retrieval from the web
    • G06F16/953Querying, e.g. by the use of web search engines
    • G06F16/9535Search customisation based on user profiles and personalisation

Definitions

  • the present application relates to the field of computer information technology, and in particular, to a project recommendation method, an electronic device, and a computer readable storage medium.
  • the present application proposes a project recommendation method, an electronic device, and a computer readable storage medium, which improves the effectiveness of the recommendation result by reconstructing the user relationship and recommending the item to a specific user based on the reconstructed user relationship. Accuracy.
  • the present application provides an electronic device including a memory and a processor, wherein the memory stores an item recommendation system operable on the processor, and the item recommendation system is The processor implements the following steps when executed:
  • a specified number of the second type of associated users are selected as the extended users of the specific users, and the extended users of the specific users are added to the updated related users of the first type, and the obtained related users are obtained.
  • the first type of associated user after the specific user is rebuilt;
  • the specified item is recommended to the specific user.
  • the correlation coefficient of the specific user and all the first-class associated users is calculated by using a first calculation formula, and the first calculation formula is set to Equation 1:
  • Equation 1 relate(u,v) represents the correlation coefficient between user u and user v, I(u) represents the set of items scored by user u, and I(v) represents the set of items scored by user v.
  • the similarity between the specific user and all the second-type associated users is calculated by using a second calculation formula, and the second calculation formula is set to Equation 2:
  • Equation 2 Representing the cosine similarity between user a and user b, a rating vector representing user a, A scoring vector representing user b.
  • the following steps are further implemented:
  • the second type of associated user whose similarity is greater than the second preset threshold is selected as the extended user of the specific user, and the extended user of the specific user is added to the updated user of the first type of the specific user, and the specific user is reconstructed. After the first class of associated users.
  • the recommending the specified item to the specific user comprises:
  • the optimized user-item scoring matrix is used to perform collaborative filtering recommendation to obtain a specified item recommended to the specific user.
  • the present application further provides a project recommendation method, which is applied to an electronic device, and the method includes:
  • a specified number of the second type of associated users are selected as the extended users of the specific users, and the extended users of the specific users are added to the updated related users of the first type, and the obtained related users are obtained.
  • the first type of associated user after the specific user is rebuilt;
  • the specified item is recommended to the specific user.
  • the correlation coefficient of the specific user and all the first-class associated users is calculated by using a first calculation formula, and the first calculation formula is set to Equation 1:
  • Equation 1 relate(u,v) represents the correlation coefficient between user u and user v, I(u) represents the set of items scored by user u, and I(v) represents the set of items scored by user v;
  • Equation 2 The similarity between the specific user and all the second-type associated users is calculated by using a second calculation formula, and the second calculation formula is set to Equation 2:
  • Equation 2 Representing the cosine similarity between user a and user b, a rating vector representing user a, A scoring vector representing user b.
  • the method further comprises:
  • the second type of associated user whose similarity is greater than the second preset threshold is selected as the extended user of the specific user, and the extended user of the specific user is added to the updated user of the first type of the specific user, and the specific user is reconstructed. After the first class of associated users.
  • the recommending the specified item to the specific user comprises:
  • the optimized user-item scoring matrix is used to perform collaborative filtering recommendation to obtain a specified item recommended to the specific user.
  • the present application further provides a computer readable storage medium storing an item recommendation system, the item recommendation system being executable by at least one processor to enable the At least one processor performs the steps of the project recommendation method as described above.
  • the electronic device, the project recommendation method, and the computer readable storage medium proposed by the present application improve the recommendation result by reconstructing the user relationship and recommending the item to the specific user based on the reconstructed user relationship.
  • Validity and accuracy, and the reconstructed user relationship avoids data sparsity issues.
  • 1 is a schematic diagram of an optional hardware architecture of an electronic device of the present application
  • FIG. 2 is a schematic diagram of a program module of an embodiment of an item recommendation system in an electronic device of the present application
  • FIG. 3 is a schematic diagram of an implementation process of an embodiment of a method for recommending an item of the present application.
  • FIG. 1 it is a schematic diagram of an optional hardware architecture of the electronic device 2 of the present application.
  • the electronic device 2 may include, but is not limited to, a memory 21, a processor 22, and a network interface 23 that can communicate with each other through a system bus. It is pointed out that FIG. 1 only shows the electronic device 2 with the components 21-23, but it should be understood that not all illustrated components are required to be implemented, and more or fewer components may be implemented instead.
  • the electronic device 2 may be a computing device such as a rack server, a blade server, a tower server, or a rack server.
  • the electronic device 2 may be an independent server or a server cluster composed of multiple servers. .
  • the memory 21 includes at least one type of readable storage medium including a flash memory, a hard disk, a multimedia card, a card type memory (eg, SD or DX memory, etc.), a random access memory (RAM), a static Random access memory (SRAM), read only memory (ROM), electrically erasable programmable read only memory (EEPROM), programmable read only memory (PROM), magnetic memory, magnetic disk, optical disk, and the like.
  • the memory 21 may be an internal storage unit of the electronic device 2, such as a hard disk or memory of the electronic device 2.
  • the memory 21 may also be an external storage device of the electronic device 2, such as a plug-in hard disk equipped on the electronic device 2, a smart memory card (SMC), and a secure digital device. (Secure Digital, SD) card, flash card, etc.
  • the memory 21 may also include both an internal storage unit of the electronic device 2 and an external storage device thereof.
  • the memory 21 is generally used to store an operating system installed in the electronic device 2 and various types of application software, such as program codes of the item recommendation system 20, and the like. Further, the memory 21 can also be used to temporarily store various types of data that have been output or are to be output.
  • the processor 22 may be a Central Processing Unit (CPU), controller, microcontroller, microprocessor, or other data processing chip in some embodiments.
  • the processor 22 is typically used to control the overall operation of the electronic device 2, such as performing control and processing related to data interaction or communication with the electronic device 2.
  • the processor 22 is configured to run program code or process data stored in the memory 21, such as running the item recommendation system 20 and the like.
  • the network interface 23 may comprise a wireless network interface or a wired network interface, which is typically used to establish a communication connection between the electronic device 2 and other electronic devices.
  • the network interface 23 is configured to connect the electronic device 2 to an external data platform through a network, and establish a data transmission channel and a communication connection between the electronic device 2 and an external data platform.
  • the network may be an intranet, an Internet, a Global System of Mobile communication (GSM), a Wideband Code Division Multiple Access (WCDMA), a 4G network, or a 5G network.
  • Wireless or wired networks such as network, Bluetooth, Wi-Fi, etc.
  • the item recommendation system 20 may be divided into one or more program modules, the one or more program modules being stored in the memory 21 and being processed by one or more processors (this The embodiment is executed by the processor 22) to complete the application.
  • the item recommendation system 20 can be divided into a calculation module 201, a screening module 202, and a recommendation module 203.
  • the program module referred to in the present application refers to a series of computer program instruction segments capable of performing a specific function, and is more suitable than the program to describe the execution process of the item recommendation system 20 in the electronic device 2. The function of each program module 201-203 will be described in detail below.
  • the calculation module 201 is configured to acquire all first-class associated users within a specified user-specific range (such as a circle of friends), and calculate a correlation coefficient between the specific user and all first-class associated users.
  • the first type of associated users in the specified user-specific range may be: all friends in the social network of the specific user, such as all friends in the circle of friends.
  • the correlation coefficient of the specific user and all the first-class associated users is calculated by using a first calculation formula, wherein the first calculation formula may be set as shown in the following formula 1.
  • replace(u,v) represents the correlation coefficient between user u and user v
  • I(u) represents the set of items scored by user u
  • I(v) represents the set of items scored by user v.
  • the screening module 202 is configured to filter out a first type of associated users whose correlation coefficient is less than a first preset threshold (for example, 60%), and associate the filtered first type of associated users from all first classes of the specific user. Deleted by the user, the first type of associated user after the specific user is updated.
  • a first preset threshold for example, 60%
  • the filtered first type of associated user represents a user who is mistakenly added without having a friend.
  • users who have similar comments on a movie if their correlation coefficient is less than the first preset threshold, means that they do not have the same interests in the movie, not the associated users with good recommendation.
  • the present application can solve the problem of data redundancy and improve the effectiveness and accuracy of subsequent project recommendation.
  • the calculation module 201 is further configured to acquire all second-class associated users outside the specified range of the specific user, and calculate similarities (such as cosine similarity) between the specific user and all second-class associated users.
  • the second type of associated users outside the specified range of the specific user may be: all the friends in the specific user address book that are not added to the circle of friends, that is, all the address book friends outside the circle of friends.
  • the similarity of the specific user with all the second type of associated users is calculated using cosine similarity.
  • the cosine similarity of the specific user and all the second-type associated users is calculated by using a second calculation formula, wherein the second calculation formula may be set as shown in the following formula 2.
  • the screening module 202 is further configured to select, according to the ranking of the similarity from high to low, a specified number of the second type of associated users (such as the top 3 second-class related users with higher similarity) as the extension of the specific user.
  • the user or potential user adds the extended user of the specific user to the first type of associated user after the specific user is updated, and obtains the first type of associated user (ie, user relationship reconstruction) after the specific user is rebuilt.
  • the data sparsity problem can be avoided by expanding the first type of associated users (associated users within the specified range) of the specific user.
  • the screening module 202 is further configured to:
  • the second type of associated user whose similarity is greater than the second preset threshold (for example, 50%) is selected as the extended user (or potential user) of the specific user, and the extended user of the specific user is added to the updated version of the specific user.
  • a type of associated user obtains the first type of associated user after the specific user is rebuilt.
  • the recommendation module 203 is configured to recommend a specified item to the specific user according to the first type of associated user after the specific user is rebuilt.
  • the recommending step specifically includes:
  • the optimized user-item scoring matrix adopts a two-dimensional matrix, each row of the two-dimensional matrix Representing one of the first type of associated users (including the specific user itself) after the reconstruction of the specific user, each column represents an item, and the two-dimensional matrix stores a rating between all users and all items, the item is Item, which represents an indicator that calculates the similarity of the user, such as a movie or song (ie, the same hobby);
  • the optimized user-item scoring matrix is used to perform collaborative filtering recommendation to obtain a specified item recommended to the specific user.
  • the traditional collaborative filtering recommendation method may be used for project recommendation, such as user-based collaborative filtering recommendation and project-based collaborative filtering recommendation.
  • the optimized user-item scoring matrix will not include the friends in the circle of friends list as long as they have the same hobbies, and there is a friend relationship in the friend circle list but there is no similar interest.
  • the user is deleted, thereby avoiding the problem that the scoring matrix is sparse, and adding relevant users to the scoring matrix, which greatly improves the accuracy of the recommendation.
  • the project recommendation system 20 proposed by the present application improves the validity and accuracy of the recommendation result by reconstructing the user relationship and recommending the project to the specific user based on the reconstructed user relationship, and The reconstructed user relationship avoids data sparsity issues.
  • the present application also proposes a project recommendation method.
  • FIG. 3 it is a schematic flowchart of an implementation process of an embodiment of the recommended method of the present application.
  • the order of execution of the steps in the flowchart shown in Fig. 3 may be changed according to different requirements, and some steps may be omitted.
  • Step S31 Acquire all first-class associated users within a specified user-specific range (such as within a circle of friends), and calculate a correlation coefficient between the specific user and all first-class associated users.
  • the first type of associated users in the specified user-specific range may be: all friends in the social network of the specific user, such as all friends in the circle of friends.
  • the correlation coefficient of the specific user and all the first-class associated users is calculated by using a first calculation formula, wherein the first calculation formula may be set as shown in the following formula 1.
  • replace(u,v) represents the correlation coefficient between user u and user v
  • I(u) represents the set of items scored by user u
  • I(v) represents the set of items scored by user v.
  • Step S32 filtering out the first type of associated users whose correlation coefficient is smaller than the first preset threshold (for example, 60%), and deleting the selected first-class associated users from all the first-class associated users of the specific user, and obtaining The first type of associated user after the specific user is updated.
  • the first preset threshold for example, 60%
  • the filtered first type of associated user represents a user who is mistakenly added without having a friend.
  • users who have similar comments on a movie if their correlation coefficient is less than the first preset threshold, means that they do not have the same interests in the movie, not the associated users with good recommendation effect.
  • the present application can solve the problem of data redundancy and improve the effectiveness and accuracy of subsequent project recommendation.
  • Step S33 Acquire all second-class associated users outside the specified range of the specific user, and calculate similarities (such as cosine similarity) between the specific user and all second-class associated users.
  • the second type of associated users outside the specified range of the specific user may be: all the friends in the specific user address book that are not added to the circle of friends, that is, all the address book friends outside the circle of friends.
  • the similarity of the particular user to all of the second type of associated users is calculated using cosine similarity.
  • the cosine similarity of the specific user and all the second-type associated users is calculated by using a second calculation formula, wherein the second calculation formula may be set as shown in the following formula 2.
  • Step S34 selecting a specified number of second-class associated users (such as the top 3 second-class associated users with higher similarity) as the extended users (or potential users) of the specific user according to the ranking of the similarity from highest to lowest.
  • the extended user of the specific user is added to the first type of associated user after the specific user is updated, and the first type of associated user (ie, user relationship reconstruction) after the specific user is rebuilt is obtained.
  • the data sparsity problem can be avoided by expanding the first type of associated users (associated users within the specified range) of the specific user.
  • step S34 may also be as follows:
  • the second type of associated user whose similarity is greater than the second preset threshold (for example, 50%) is selected as the extended user (or potential user) of the specific user, and the extended user of the specific user is added to the updated version of the specific user.
  • a type of associated user obtains the first type of associated user after the specific user is rebuilt.
  • Step S35 recommending a specified item to the specific user according to the first type of associated user after the specific user is rebuilt.
  • the recommending step specifically includes:
  • the optimized user-item scoring matrix adopts a two-dimensional matrix, each row of the two-dimensional matrix Representing one of the first type of associated users (including the specific user itself) after the reconstruction of the specific user, each column represents an item, and the two-dimensional matrix stores a rating between all users and all items, the item is Item, which represents an indicator that calculates the similarity of the user, such as a movie or song (ie, the same hobby);
  • the optimized user-item scoring matrix is used to perform collaborative filtering recommendation to obtain a specified item recommended to the specific user.
  • the traditional collaborative filtering recommendation method may be used for project recommendation, such as user-based collaborative filtering recommendation and project-based collaborative filtering recommendation.
  • the optimized user-item scoring matrix will not include the friends in the circle of friends list as long as they have the same hobbies, and there is a friend relationship in the friend circle list but there is no similar interest.
  • the user is deleted, thereby avoiding the problem that the scoring matrix is sparse, and adding relevant users to the scoring matrix, which greatly improves the accuracy of the recommendation.
  • the project recommendation method proposed by the present application improves the validity and accuracy of the recommendation result by reconstructing the user relationship and recommending the project to a specific user based on the reconstructed user relationship. And the reconstructed user relationship avoids the problem of data sparsity.
  • the present application further provides a computer readable storage medium (such as a ROM/RAM, a magnetic disk, an optical disk), the computer readable storage medium storing an item recommendation system 20, the item recommendation system 20 may be executed by at least one processor 22 to cause the at least one processor 22 to perform the steps of the item recommendation method as described above.
  • a computer readable storage medium such as a ROM/RAM, a magnetic disk, an optical disk
  • the item recommendation system 20 may be executed by at least one processor 22 to cause the at least one processor 22 to perform the steps of the item recommendation method as described above.
  • the foregoing embodiment method can be implemented by means of software plus a necessary general hardware platform, and can also be implemented by hardware, but in many cases, the former is A better implementation.
  • the technical solution of the present application which is essential or contributes to the prior art, may be embodied in the form of a software product stored in a storage medium (such as ROM/RAM, disk,
  • the optical disc includes a number of instructions for causing a terminal device (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to perform the methods described in various embodiments of the present application.

Landscapes

  • Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Primary Health Care (AREA)
  • Marketing (AREA)
  • Human Resources & Organizations (AREA)
  • Strategic Management (AREA)
  • Tourism & Hospitality (AREA)
  • General Health & Medical Sciences (AREA)
  • General Business, Economics & Management (AREA)
  • Economics (AREA)
  • Health & Medical Sciences (AREA)
  • Computing Systems (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

本申请公开了一种项目推荐方法,该方法包括步骤:计算特定用户与第一类关联用户的相关系数;筛选出相关系数小于第一预设阈值的第一类关联用户并删除,得到该特定用户更新后的第一类关联用户;计算该特定用户与第二类关联用户的相似度;依据相似度大小选取指定数量的第二类关联用户作为该特定用户的拓展用户,将该特定用户的拓展用户添加至该特定用户更新后的第一类关联用户,得到该特定用户重建后的第一类关联用户;根据该特定用户重建后的第一类关联用户,向该特定用户推荐指定项目。本申请可以提高推荐结果的有效性和准确率。

Description

项目推荐方法、电子设备及计算机可读存储介质
本申请要求于2017年11月01日提交中国专利局、申请号为201711059170.7、发明名称为“项目推荐方法、电子设备及计算机可读存储介质”的中国专利申请的优先权,其全部内容通过引用结合在申请中。
技术领域
本申请涉及计算机信息技术领域,尤其涉及一种项目推荐方法、电子设备及计算机可读存储介质。
背景技术
目前,大多数基于用户关系的项目推荐方案都是直接采集用户关系数据进行推荐。但是,在社交网络中建立朋友关系是非常自由的,很容易出现朋友关系不可靠的现象,导致大量的冗余数据。这些大量的冗余数据直接影响了推荐结果的有效性和准确性。而对用户关系数据进行简单的过滤处理又很容易引起数据稀疏性等问题。故,现有技术中的项目推荐方法设计不够合理,亟需改进。
发明内容
有鉴于此,本申请提出一种项目推荐方法、电子设备及计算机可读存储介质,通过对用户关系进行重建,并基于重建后的用户关系向特定用户推荐项目,提高了推荐结果的有效性和准确率。
首先,为实现上述目的,本申请提出一种电子设备,所述电子设备包括存储器及处理器,所述存储器上存储有可在所述处理器上运行的项目推荐系统,所述项目推荐系统被所述处理器执行时实现如下步骤:
获取特定用户指定范围内的所有第一类关联用户,计算该特定用户与所 有第一类关联用户的相关系数;
筛选出相关系数小于第一预设阈值的第一类关联用户,并将筛选出的第一类关联用户从该特定用户的所有第一类关联用户中删除,得到该特定用户更新后的第一类关联用户;
获取该特定用户指定范围外的所有第二类关联用户,计算该特定用户与所有第二类关联用户的相似度;
依据相似度从高至低的排序,选取指定数量的第二类关联用户作为该特定用户的拓展用户,将该特定用户的拓展用户添加至该特定用户更新后的第一类关联用户,得到该特定用户重建后的第一类关联用户;及
根据该特定用户重建后的第一类关联用户,向该特定用户推荐指定项目。
优选地,该特定用户与所有第一类关联用户的相关系数采用第一计算公式计算得出,所述第一计算公式设置为公式1:
Figure PCTCN2018076171-appb-000001
公式1中,relate(u,v)代表用户u与用户v的相关系数,I(u)代表用户u评分过的项目集合,I(v)代表用户v评分过的项目集合。
优选地,该特定用户与所有第二类关联用户的相似度采用第二计算公式计算得出,所述第二计算公式设置为公式2:
Figure PCTCN2018076171-appb-000002
公式2中,
Figure PCTCN2018076171-appb-000003
代表用户a和用户b的余弦相似度,
Figure PCTCN2018076171-appb-000004
代表用户a的评分向量,
Figure PCTCN2018076171-appb-000005
代表用户b的评分向量。
优选地,所述项目推荐系统被所述处理器执行时还用于实现如下步骤:
筛选出相似度大于第二预设阈值的第二类关联用户作为该特定用户的拓展用户,将该特定用户的拓展用户添加至该特定用户更新后的第一类关联用户,得到该特定用户重建后的第一类关联用户。
优选地,所述向该特定用户推荐指定项目包括:
根据该特定用户重建后的第一类关联用户,构建该特定用户优化后的用户-项目评分矩阵,其中,所述优化后的用户-项目评分矩阵采用二维矩阵,该二维矩阵的每一行代表该特定用户重建后的第一类关联用户中的一个用户,每一列代表一个项目,且该二维矩阵存储有所有用户与所有项目之间的评分;
利用所述优化后的用户-项目评分矩阵进行协同过滤推荐,得到向该特定用户推荐的指定项目。
此外,为实现上述目的,本申请还提供一种项目推荐方法,该方法应用于电子设备,所述方法包括:
获取特定用户指定范围内的所有第一类关联用户,计算该特定用户与所有第一类关联用户的相关系数;
筛选出相关系数小于第一预设阈值的第一类关联用户,并将筛选出的第一类关联用户从该特定用户的所有第一类关联用户中删除,得到该特定用户更新后的第一类关联用户;
获取该特定用户指定范围外的所有第二类关联用户,计算该特定用户与所有第二类关联用户的相似度;
依据相似度从高至低的排序,选取指定数量的第二类关联用户作为该特定用户的拓展用户,将该特定用户的拓展用户添加至该特定用户更新后的第一类关联用户,得到该特定用户重建后的第一类关联用户;及
根据该特定用户重建后的第一类关联用户,向该特定用户推荐指定项目。
优选地,该特定用户与所有第一类关联用户的相关系数采用第一计算公式计算得出,所述第一计算公式设置为公式1:
Figure PCTCN2018076171-appb-000006
公式1中,relate(u,v)代表用户u与用户v的相关系数,I(u)代表用户u评分过的项目集合,I(v)代表用户v评分过的项目集合;
该特定用户与所有第二类关联用户的相似度采用第二计算公式计算得 出,所述第二计算公式设置为公式2:
Figure PCTCN2018076171-appb-000007
公式2中,
Figure PCTCN2018076171-appb-000008
代表用户a和用户b的余弦相似度,
Figure PCTCN2018076171-appb-000009
代表用户a的评分向量,
Figure PCTCN2018076171-appb-000010
代表用户b的评分向量。
优选地,该方法还包括:
筛选出相似度大于第二预设阈值的第二类关联用户作为该特定用户的拓展用户,将该特定用户的拓展用户添加至该特定用户更新后的第一类关联用户,得到该特定用户重建后的第一类关联用户。
优选地,所述向该特定用户推荐指定项目包括:
根据该特定用户重建后的第一类关联用户,构建该特定用户优化后的用户-项目评分矩阵,其中,所述优化后的用户-项目评分矩阵采用二维矩阵,该二维矩阵的每一行代表该特定用户重建后的第一类关联用户中的一个用户,每一列代表一个项目,且该二维矩阵存储有所有用户与所有项目之间的评分;
利用所述优化后的用户-项目评分矩阵进行协同过滤推荐,得到向该特定用户推荐的指定项目。
进一步地,为实现上述目的,本申请还提供一种计算机可读存储介质,所述计算机可读存储介质存储有项目推荐系统,所述项目推荐系统可被至少一个处理器执行,以使所述至少一个处理器执行如上述的项目推荐方法的步骤。
相较于现有技术,本申请所提出的电子设备、项目推荐方法及计算机可读存储介质,通过对用户关系进行重建,并基于重建后的用户关系向特定用户推荐项目,提高了推荐结果的有效性和准确率,且重建后的用户关系避免了数据稀疏性问题。
附图说明
图1是本申请电子设备一可选的硬件架构的示意图;
图2是本申请电子设备中项目推荐系统一实施例的程序模块示意图;
图3为本申请项目推荐方法一实施例的实施流程示意图。
附图标记:
电子设备 2
存储器 21
处理器 22
网络接口 23
项目推荐系统 20
计算模块 201
筛选模块 202
推荐模块 203
流程步骤 S31-S35
本申请目的的实现、功能特点及优点将结合实施例,参照附图做进一步说明。
具体实施方式
为了使本申请的目的、技术方案及优点更加清楚明白,以下结合附图及实施例,对本申请进行进一步详细说明。应当理解,此处所描述的具体实施例仅用以解释本申请,并不用于限定本申请。基于本申请中的实施例,本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例,都属于本申请保护的范围。
需要说明的是,在本申请中涉及“第一”、“第二”等的描述仅用于描述 目的,而不能理解为指示或暗示其相对重要性或者隐含指明所指示的技术特征的数量。由此,限定有“第一”、“第二”的特征可以明示或者隐含地包括至少一个该特征。另外,各个实施例之间的技术方案可以相互结合,但是必须是以本领域普通技术人员能够实现为基础,当技术方案的结合出现相互矛盾或无法实现时应当认为这种技术方案的结合不存在,也不在本申请要求的保护范围之内。
进一步需要说明的是,在本文中,术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含,从而使得包括一系列要素的过程、方法、物品或者装置不仅包括那些要素,而且还包括没有明确列出的其他要素,或者是还包括为这种过程、方法、物品或者装置所固有的要素。在没有更多限制的情况下,由语句“包括一个……”限定的要素,并不排除在包括该要素的过程、方法、物品或者装置中还存在另外的相同要素。
首先,本申请提出一种电子设备2。
参阅图1所示,是本申请电子设备2一可选的硬件架构的示意图。本实施例中,所述电子设备2可包括,但不限于,可通过系统总线相互通信连接存储器21、处理器22、网络接口23。需要指出的是,图1仅示出了具有组件21-23的电子设备2,但是应理解的是,并不要求实施所有示出的组件,可以替代的实施更多或者更少的组件。
其中,所述电子设备2可以是机架式服务器、刀片式服务器、塔式服务器或机柜式服务器等计算设备,该电子设备2可以是独立的服务器,也可以是多个服务器所组成的服务器集群。
所述存储器21至少包括一种类型的可读存储介质,所述可读存储介质包括闪存、硬盘、多媒体卡、卡型存储器(例如,SD或DX存储器等)、随机访问存储器(RAM)、静态随机访问存储器(SRAM)、只读存储器(ROM)、电可擦除可编程只读存储器(EEPROM)、可编程只读存储器(PROM)、磁性 存储器、磁盘、光盘等。在一些实施例中,所述存储器21可以是所述电子设备2的内部存储单元,例如该电子设备2的硬盘或内存。在另一些实施例中,所述存储器21也可以是所述电子设备2的外部存储设备,例如该电子设备2上配备的插接式硬盘,智能存储卡(Smart Media Card,SMC),安全数字(Secure Digital,SD)卡,闪存卡(Flash Card)等。当然,所述存储器21还可以既包括所述电子设备2的内部存储单元也包括其外部存储设备。本实施例中,所述存储器21通常用于存储安装于所述电子设备2的操作系统和各类应用软件,例如所述项目推荐系统20的程序代码等。此外,所述存储器21还可以用于暂时地存储已经输出或者将要输出的各类数据。
所述处理器22在一些实施例中可以是中央处理器(Central Processing Unit,CPU)、控制器、微控制器、微处理器、或其他数据处理芯片。该处理器22通常用于控制所述电子设备2的总体操作,例如执行与所述电子设备2进行数据交互或者通信相关的控制和处理等。本实施例中,所述处理器22用于运行所述存储器21中存储的程序代码或者处理数据,例如运行所述的项目推荐系统20等。
所述网络接口23可包括无线网络接口或有线网络接口,该网络接口23通常用于在所述电子设备2与其他电子设备之间建立通信连接。例如,所述网络接口23用于通过网络将所述电子设备2与外部数据平台相连,在所述电子设备2与外部数据平台之间的建立数据传输通道和通信连接。所述网络可以是企业内部网(Intranet)、互联网(Internet)、全球移动通讯系统(Global System of Mobile communication,GSM)、宽带码分多址(Wideband Code Division Multiple Access,WCDMA)、4G网络、5G网络、蓝牙(Bluetooth)、Wi-Fi等无线或有线网络。
至此,己经详细介绍了本申请各个实施例的应用环境和相关设备的硬件结构和功能。下面,将基于上述应用环境和相关设备,提出本申请的各个实施例。
参阅图2所示,是本申请电子设备2中项目推荐系统20一实施例的程序模块图。本实施例中,所述的项目推荐系统20可以被分割成一个或多个程序模块,所述一个或者多个程序模块被存储于所述存储器21中,并由一个或多个处理器(本实施例中为所述处理器22)所执行,以完成本申请。例如,在图2中,所述的项目推荐系统20可以被分割成计算模块201、筛选模块202、以及推荐模块203。本申请所称的程序模块是指能够完成特定功能的一系列计算机程序指令段,比程序更适合于描述所述项目推荐系统20在所述电子设备2中的执行过程。以下将就各程序模块201-203的功能进行详细描述。
所述计算模块201,用于获取特定用户指定范围内(如朋友圈内)的所有第一类关联用户,计算该特定用户与所有第一类关联用户的相关系数。其中,该特定用户指定范围内的所有第一类关联用户可以是:该特定用户社交网络中的所有好友,如朋友圈中的所有好友。
优选地,在本实施例中,该特定用户与所有第一类关联用户的相关系数采用第一计算公式计算得出,其中,所述第一计算公式可以设置为如下公式1所示。
Figure PCTCN2018076171-appb-000011
其中,relate(u,v)代表用户u与用户v的相关系数,I(u)代表用户u评分过的项目集合,I(v)代表用户v评分过的项目集合。
所述筛选模块202,用于筛选出相关系数小于第一预设阈值(如60%)的第一类关联用户,并将筛选出的第一类关联用户从该特定用户的所有第一类关联用户中删除,得到该特定用户更新后的第一类关联用户。
优选地,在本实施例中,所述筛选出的第一类关联用户代表不具备朋友资格而被误加的用户。例如,对某个电影有相似评论的用户,如果他们的相关系数小于第一预设阈值,则代表他们在电影上并没有相同兴趣爱好,不是 具有良好推荐效果的关联用户。本申请通过删除相关系数较小的关联用户,可以解决数据冗余问题,提高后续项目推荐的有效性和准确率。
所述计算模块201,还用于获取该特定用户指定范围外的所有第二类关联用户,计算该特定用户与所有第二类关联用户的相似度(如余弦相似度)。其中,该特定用户指定范围外的所有第二类关联用户可以是:该特定用户通讯录中没有加入朋友圈的所有好友,即朋友圈外的所有通讯录好友。
优选地,在本实施例中,该特定用户与所有第二类关联用户的相似度采用余弦相似度进行计算。进一步地,该特定用户与所有第二类关联用户的余弦相似度采用第二计算公式计算得出,其中,所述第二计算公式可以设置为如下公式2所示。
Figure PCTCN2018076171-appb-000012
其中,
Figure PCTCN2018076171-appb-000013
代表用户a和用户b的余弦相似度,
Figure PCTCN2018076171-appb-000014
代表用户a的评分向量,
Figure PCTCN2018076171-appb-000015
代表用户b的评分向量。
所述筛选模块202,还用于依据相似度从高至低的排序,选取指定数量的第二类关联用户(如相似度较高的前3位第二类关联用户)作为该特定用户的拓展用户(或潜在用户),将该特定用户的拓展用户添加至该特定用户更新后的第一类关联用户,得到该特定用户重建后的第一类关联用户(即用户关系重建)。在本实施例中,通过对该特定用户的第一类关联用户(指定范围内的关联用户)进行拓展,可以避免数据稀疏性问题。
需要说明的是,在其它实施例中,所述筛选模块202还用于:
筛选出相似度大于第二预设阈值(如50%)的第二类关联用户作为该特定用户的拓展用户(或潜在用户),将该特定用户的拓展用户添加至该特定用户 更新后的第一类关联用户,得到该特定用户重建后的第一类关联用户。
所述推荐模块203,用于根据该特定用户重建后的第一类关联用户,向该特定用户推荐指定项目。
优选地,在本实施例中,推荐步骤具体包括:
根据该特定用户重建后的第一类关联用户,构建该特定用户优化后的用户-项目评分矩阵,其中,所述优化后的用户-项目评分矩阵采用二维矩阵,该二维矩阵的每一行代表该特定用户重建后的第一类关联用户中的一个用户(包括该特定用户自己),每一列代表一个项目,且该二维矩阵存储有所有用户与所有项目之间的评分,该项目为item,代表计算用户相似度的某一项指标,比如同时喜欢某个电影或者歌曲(即相同的兴趣爱好);
利用所述优化后的用户-项目评分矩阵进行协同过滤推荐,得到向该特定用户推荐的指定项目。在本实施例中,可以采用传统的协同过滤推荐方法进行项目推荐,如基于用户的协同过滤推荐和基于项目的协同过滤推荐。
在本实施例中,优化后的用户-项目评分矩阵将不在朋友圈列表的朋友也加入预测范围(只要是有相同的兴趣爱好),并且将朋友圈列表中虽然存在朋友关系但没有相同兴趣爱好的用户删除,从而避免了评分矩阵稀疏的问题,而且将相关用户加入评分矩阵,极大提高了推荐的准确度。
通过上述程序模块201-203,本申请所提出的项目推荐系统20,通过对用户关系进行重建,并基于重建后的用户关系向特定用户推荐项目,提高了推荐结果的有效性和准确率,且重建后的用户关系避免了数据稀疏性问题。
此外,本申请还提出一种项目推荐方法。
参阅图3所示,是本申请项目推荐方法一实施例的实施流程示意图。在本实施例中,根据不同的需求,图3所示的流程图中的步骤的执行顺序可以 改变,某些步骤可以省略。
步骤S31,获取特定用户指定范围内(如朋友圈内)的所有第一类关联用户,计算该特定用户与所有第一类关联用户的相关系数。其中,该特定用户指定范围内的所有第一类关联用户可以是:该特定用户社交网络中的所有好友,如朋友圈中的所有好友。
优选地,在本实施例中,该特定用户与所有第一类关联用户的相关系数采用第一计算公式计算得出,其中,所述第一计算公式可以设置为如下公式1所示。
Figure PCTCN2018076171-appb-000016
其中,relate(u,v)代表用户u与用户v的相关系数,I(u)代表用户u评分过的项目集合,I(v)代表用户v评分过的项目集合。
步骤S32,筛选出相关系数小于第一预设阈值(如60%)的第一类关联用户,并将筛选出的第一类关联用户从该特定用户的所有第一类关联用户中删除,得到该特定用户更新后的第一类关联用户。
优选地,在本实施例中,所述筛选出的第一类关联用户代表不具备朋友资格而被误加的用户。例如,对某个电影有相似评论的用户,如果他们的相关系数小于第一预设阈值,则代表他们在电影上并没有相同兴趣爱好,不是具有良好推荐效果的关联用户。本申请通过删除相关系数较小的关联用户,可以解决数据冗余问题,提高后续项目推荐的有效性和准确率。
步骤S33,获取该特定用户指定范围外的所有第二类关联用户,计算该特定用户与所有第二类关联用户的相似度(如余弦相似度)。其中,该特定用户指定范围外的所有第二类关联用户可以是:该特定用户通讯录中没有加入朋友圈的所有好友,即朋友圈外的所有通讯录好友。
优选地,在本实施例中,该特定用户与所有第二类关联用户的相似度采 用余弦相似度进行计算。进一步地,该特定用户与所有第二类关联用户的余弦相似度采用第二计算公式计算得出,其中,所述第二计算公式可以设置为如下公式2所示。
Figure PCTCN2018076171-appb-000017
其中,
Figure PCTCN2018076171-appb-000018
代表用户a和用户b的余弦相似度,
Figure PCTCN2018076171-appb-000019
代表用户a的评分向量,
Figure PCTCN2018076171-appb-000020
代表用户b的评分向量。
步骤S34,依据相似度从高至低的排序,选取指定数量的第二类关联用户(如相似度较高的前3位第二类关联用户)作为该特定用户的拓展用户(或潜在用户),将该特定用户的拓展用户添加至该特定用户更新后的第一类关联用户,得到该特定用户重建后的第一类关联用户(即用户关系重建)。在本实施例中,通过对该特定用户的第一类关联用户(指定范围内的关联用户)进行拓展,可以避免数据稀疏性问题。
需要说明的是,在其它实施例中,所述步骤S34也可以是如下步骤:
筛选出相似度大于第二预设阈值(如50%)的第二类关联用户作为该特定用户的拓展用户(或潜在用户),将该特定用户的拓展用户添加至该特定用户更新后的第一类关联用户,得到该特定用户重建后的第一类关联用户。
步骤S35,根据该特定用户重建后的第一类关联用户,向该特定用户推荐指定项目。
优选地,在本实施例中,推荐步骤具体包括:
根据该特定用户重建后的第一类关联用户,构建该特定用户优化后的用户-项目评分矩阵,其中,所述优化后的用户-项目评分矩阵采用二维矩阵,该二维矩阵的每一行代表该特定用户重建后的第一类关联用户中的一个用户 (包括该特定用户自己),每一列代表一个项目,且该二维矩阵存储有所有用户与所有项目之间的评分,该项目为item,代表计算用户相似度的某一项指标,比如同时喜欢某个电影或者歌曲(即相同的兴趣爱好);
利用所述优化后的用户-项目评分矩阵进行协同过滤推荐,得到向该特定用户推荐的指定项目。在本实施例中,可以采用传统的协同过滤推荐方法进行项目推荐,如基于用户的协同过滤推荐和基于项目的协同过滤推荐。
在本实施例中,优化后的用户-项目评分矩阵将不在朋友圈列表的朋友也加入预测范围(只要是有相同的兴趣爱好),并且将朋友圈列表中虽然存在朋友关系但没有相同兴趣爱好的用户删除,从而避免了评分矩阵稀疏的问题,而且将相关用户加入评分矩阵,极大提高了推荐的准确度。
通过上述步骤S31-S35及其它相关步骤,本申请所提出的项目推荐方法,通过对用户关系进行重建,并基于重建后的用户关系向特定用户推荐项目,提高了推荐结果的有效性和准确率,且重建后的用户关系避免了数据稀疏性问题。
进一步地,为实现上述目的,本申请还提供一种计算机可读存储介质(如ROM/RAM、磁碟、光盘),所述计算机可读存储介质存储有项目推荐系统20,所述项目推荐系统20可被至少一个处理器22执行,以使所述至少一个处理器22执行如上所述的项目推荐方法的步骤。
通过以上的实施方式的描述,本领域的技术人员可以清楚地了解到上述实施例方法可借助软件加必需的通用硬件平台的方式来实现,当然也可以通过硬件来实现,但很多情况下前者是更佳的实施方式。基于这样的理解,本申请的技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来,该计算机软件产品存储在一个存储介质(如ROM/RAM、磁碟、 光盘)中,包括若干指令用以使得一台终端设备(可以是手机,计算机,服务器,空调器,或者网络设备等)执行本申请各个实施例所述的方法。
以上参照附图说明了本申请的优选实施例,并非因此局限本申请的权利范围。上述本申请实施例序号仅仅为了描述,不代表实施例的优劣。另外,虽然在流程图中示出了逻辑顺序,但是在某些情况下,可以以不同于此处的顺序执行所示出或描述的步骤。
本领域技术人员不脱离本申请的范围和实质,可以有多种变型方案实现本申请,比如作为一个实施例的特征可用于另一实施例而得到又一实施例。凡是利用本申请说明书及附图内容所作的等效结构或等效流程变换,或直接或间接运用在其他相关的技术领域,均同理包括在本申请的专利保护范围内。

Claims (20)

  1. 一种电子设备,其特征在于,所述电子设备包括存储器及处理器,所述存储器上存储有可在所述处理器上运行的项目推荐系统,所述项目推荐系统被所述处理器执行时实现如下步骤:
    获取特定用户指定范围内的所有第一类关联用户,计算该特定用户与所有第一类关联用户的相关系数;
    筛选出相关系数小于第一预设阈值的第一类关联用户,并将筛选出的第一类关联用户从该特定用户的所有第一类关联用户中删除,得到该特定用户更新后的第一类关联用户;
    获取该特定用户指定范围外的所有第二类关联用户,计算该特定用户与所有第二类关联用户的相似度;
    依据相似度从高至低的排序,选取指定数量的第二类关联用户作为该特定用户的拓展用户,将该特定用户的拓展用户添加至该特定用户更新后的第一类关联用户,得到该特定用户重建后的第一类关联用户;及
    根据该特定用户重建后的第一类关联用户,向该特定用户推荐指定项目。
  2. 如权利要求1所述的电子设备,其特征在于,该特定用户与所有第一类关联用户的相关系数采用第一计算公式计算得出,所述第一计算公式设置为公式1:
    Figure PCTCN2018076171-appb-100001
    公式1中,relate(u,v)代表用户u与用户v的相关系数,I(u)代表用户u评分过的项目集合,I(v)代表用户v评分过的项目集合。
  3. 如权利要求1所述的电子设备,其特征在于,该特定用户与所有第二类关联用户的相似度采用第二计算公式计算得出,所述第二计算公式设置为 公式2:
    Figure PCTCN2018076171-appb-100002
    公式2中,
    Figure PCTCN2018076171-appb-100003
    代表用户a和用户b的余弦相似度,
    Figure PCTCN2018076171-appb-100004
    代表用户a的评分向量,
    Figure PCTCN2018076171-appb-100005
    代表用户b的评分向量。
  4. 如权利要求1所述的电子设备,其特征在于,所述项目推荐系统被所述处理器执行时还用于实现如下步骤:
    筛选出相似度大于第二预设阈值的第二类关联用户作为该特定用户的拓展用户,将该特定用户的拓展用户添加至该特定用户更新后的第一类关联用户,得到该特定用户重建后的第一类关联用户。
  5. 如权利要求2所述的电子设备,其特征在于,所述项目推荐系统被所述处理器执行时还用于实现如下步骤:
    筛选出相似度大于第二预设阈值的第二类关联用户作为该特定用户的拓展用户,将该特定用户的拓展用户添加至该特定用户更新后的第一类关联用户,得到该特定用户重建后的第一类关联用户。
  6. 如权利要求3所述的电子设备,其特征在于,所述项目推荐系统被所述处理器执行时还用于实现如下步骤:
    筛选出相似度大于第二预设阈值的第二类关联用户作为该特定用户的拓展用户,将该特定用户的拓展用户添加至该特定用户更新后的第一类关联用户,得到该特定用户重建后的第一类关联用户。
  7. 如权利要求1-6任一项所述的电子设备,其特征在于,所述向该特定用户推荐指定项目包括:
    根据该特定用户重建后的第一类关联用户,构建该特定用户优化后的用户-项目评分矩阵,其中,所述优化后的用户-项目评分矩阵采用二维矩阵,该二维矩阵的每一行代表该特定用户重建后的第一类关联用户中的一个用户,每一列代表一个项目,且该二维矩阵存储有所有用户与所有项目之间的评分;
    利用所述优化后的用户-项目评分矩阵进行协同过滤推荐,得到向该特定用户推荐的指定项目。
  8. 一种项目推荐方法,应用于电子设备,其特征在于,所述方法包括:
    获取特定用户指定范围内的所有第一类关联用户,计算该特定用户与所有第一类关联用户的相关系数;
    筛选出相关系数小于第一预设阈值的第一类关联用户,并将筛选出的第一类关联用户从该特定用户的所有第一类关联用户中删除,得到该特定用户更新后的第一类关联用户;
    获取该特定用户指定范围外的所有第二类关联用户,计算该特定用户与所有第二类关联用户的相似度;
    依据相似度从高至低的排序,选取指定数量的第二类关联用户作为该特定用户的拓展用户,将该特定用户的拓展用户添加至该特定用户更新后的第一类关联用户,得到该特定用户重建后的第一类关联用户;及
    根据该特定用户重建后的第一类关联用户,向该特定用户推荐指定项目。
  9. 如权利要求8所述的项目推荐方法,其特征在于,该特定用户与所有第一类关联用户的相关系数采用第一计算公式计算得出,所述第一计算公式设置为公式1:
    Figure PCTCN2018076171-appb-100006
    公式1中,relate(u,v)代表用户u与用户v的相关系数,I(u)代表用户u评分过的项目集合,I(v)代表用户v评分过的项目集合。
  10. 如权利要求8所述的项目推荐方法,其特征在于,该特定用户与所有第二类关联用户的相似度采用第二计算公式计算得出,所述第二计算公式设置为公式2:
    Figure PCTCN2018076171-appb-100007
    公式2中,
    Figure PCTCN2018076171-appb-100008
    代表用户a和用户b的余弦相似度,
    Figure PCTCN2018076171-appb-100009
    代表用户a的评分向量,
    Figure PCTCN2018076171-appb-100010
    代表用户b的评分向量。
  11. 如权利要求8所述的项目推荐方法,其特征在于,该方法还包括:
    筛选出相似度大于第二预设阈值的第二类关联用户作为该特定用户的拓展用户,将该特定用户的拓展用户添加至该特定用户更新后的第一类关联用户,得到该特定用户重建后的第一类关联用户。
  12. 如权利要求9所述的项目推荐方法,其特征在于,该方法还包括:
    筛选出相似度大于第二预设阈值的第二类关联用户作为该特定用户的拓展用户,将该特定用户的拓展用户添加至该特定用户更新后的第一类关联用户,得到该特定用户重建后的第一类关联用户。
  13. 如权利要求10所述的项目推荐方法,其特征在于,该方法还包括:
    筛选出相似度大于第二预设阈值的第二类关联用户作为该特定用户的拓展用户,将该特定用户的拓展用户添加至该特定用户更新后的第一类关联用户,得到该特定用户重建后的第一类关联用户。
  14. 如权利要求8-13任一项所述的项目推荐方法,其特征在于,所述向该特定用户推荐指定项目包括:
    根据该特定用户重建后的第一类关联用户,构建该特定用户优化后的用户-项目评分矩阵,其中,所述优化后的用户-项目评分矩阵采用二维矩阵,该二维矩阵的每一行代表该特定用户重建后的第一类关联用户中的一个用户,每一列代表一个项目,且该二维矩阵存储有所有用户与所有项目之间的评分;
    利用所述优化后的用户-项目评分矩阵进行协同过滤推荐,得到向该特定用户推荐的指定项目。
  15. 一种计算机可读存储介质,所述计算机可读存储介质存储有项目推荐系统,所述项目推荐系统可被至少一个处理器执行,所述项目推荐系统被所述处理器执行时实现如下步骤:
    获取特定用户指定范围内的所有第一类关联用户,计算该特定用户与所有第一类关联用户的相关系数;
    筛选出相关系数小于第一预设阈值的第一类关联用户,并将筛选出的第一类关联用户从该特定用户的所有第一类关联用户中删除,得到该特定用户更新后的第一类关联用户;
    获取该特定用户指定范围外的所有第二类关联用户,计算该特定用户与所有第二类关联用户的相似度;
    依据相似度从高至低的排序,选取指定数量的第二类关联用户作为该特定用户的拓展用户,将该特定用户的拓展用户添加至该特定用户更新后的第一类关联用户,得到该特定用户重建后的第一类关联用户;及
    根据该特定用户重建后的第一类关联用户,向该特定用户推荐指定项目。
  16. 如权利要求15所述的计算机可读存储介质,其特征在于,该特定用户与所有第一类关联用户的相关系数采用第一计算公式计算得出,所述第一计算公式设置为公式1:
    Figure PCTCN2018076171-appb-100011
    公式1中,relate(u,v)代表用户u与用户v的相关系数,I(u)代表用户u评分过的项目集合,I(v)代表用户v评分过的项目集合。
  17. 如权利要求15所述的计算机可读存储介质,其特征在于,该特定用户与所有第二类关联用户的相似度采用第二计算公式计算得出,所述第二计算公式设置为公式2:
    Figure PCTCN2018076171-appb-100012
    公式2中,
    Figure PCTCN2018076171-appb-100013
    代表用户a和用户b的余弦相似度,
    Figure PCTCN2018076171-appb-100014
    代表用户a的评分向量,
    Figure PCTCN2018076171-appb-100015
    代表用户b的评分向量。
  18. 如权利要求15所述的计算机可读存储介质,其特征在于,所述项目推荐系统被所述处理器执行时还用于实现如下步骤:
    筛选出相似度大于第二预设阈值的第二类关联用户作为该特定用户的拓展用户,将该特定用户的拓展用户添加至该特定用户更新后的第一类关联用户,得到该特定用户重建后的第一类关联用户。
  19. 如权利要求16或17所述的计算机可读存储介质,其特征在于,所述项目推荐系统被所述处理器执行时还用于实现如下步骤:
    筛选出相似度大于第二预设阈值的第二类关联用户作为该特定用户的拓展用户,将该特定用户的拓展用户添加至该特定用户更新后的第一类关联用户,得到该特定用户重建后的第一类关联用户。
  20. 如权利要求19所述的计算机可读存储介质,其特征在于,所述向该特定用户推荐指定项目包括:
    根据该特定用户重建后的第一类关联用户,构建该特定用户优化后的用 户-项目评分矩阵,其中,所述优化后的用户-项目评分矩阵采用二维矩阵,该二维矩阵的每一行代表该特定用户重建后的第一类关联用户中的一个用户,每一列代表一个项目,且该二维矩阵存储有所有用户与所有项目之间的评分;
    利用所述优化后的用户-项目评分矩阵进行协同过滤推荐,得到向该特定用户推荐的指定项目。
PCT/CN2018/076171 2017-11-01 2018-02-10 项目推荐方法、电子设备及计算机可读存储介质 WO2019085333A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201711059170.7A CN108038789A (zh) 2017-11-01 2017-11-01 项目推荐方法、电子设备及计算机可读存储介质
CN201711059170.7 2017-11-01

Publications (1)

Publication Number Publication Date
WO2019085333A1 true WO2019085333A1 (zh) 2019-05-09

Family

ID=62093543

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/076171 WO2019085333A1 (zh) 2017-11-01 2018-02-10 项目推荐方法、电子设备及计算机可读存储介质

Country Status (2)

Country Link
CN (1) CN108038789A (zh)
WO (1) WO2019085333A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112287166B (zh) * 2020-09-23 2023-03-07 山东师范大学 一种基于改进深度信念网络的电影推荐方法及系统

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110320462A1 (en) * 2010-06-29 2011-12-29 International Business Machines Corporation Method and apparatus for recommending information to users within a social network
CN103138954A (zh) * 2011-12-02 2013-06-05 中国移动通信集团公司 一种推荐项的推送方法、系统及推荐服务器
CN103428169A (zh) * 2012-05-17 2013-12-04 腾讯科技(深圳)有限公司 Sns社区中推荐用户的方法和系统
CN103678518A (zh) * 2013-11-28 2014-03-26 北京邮电大学 一种推荐列表调整方法和装置
CN104202319A (zh) * 2014-08-28 2014-12-10 北京淘友天下科技发展有限公司 一种社交关系推荐方法及装置
CN104615609A (zh) * 2014-04-30 2015-05-13 腾讯科技(深圳)有限公司 一种联系人管理方法及装置

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103377250B (zh) * 2012-04-27 2017-08-04 杭州载言网络技术有限公司 基于邻域的top‑k推荐方法
US20150379609A1 (en) * 2014-06-30 2015-12-31 Kobo Incorporated Generating recommendations for unfamiliar users by utilizing social side information
CN105608179B (zh) * 2015-12-22 2019-03-12 百度在线网络技术(北京)有限公司 确定用户标识的关联性的方法和装置
CN106649540B (zh) * 2016-10-26 2022-04-01 Tcl科技集团股份有限公司 一种视频推荐方法及系统

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110320462A1 (en) * 2010-06-29 2011-12-29 International Business Machines Corporation Method and apparatus for recommending information to users within a social network
CN103138954A (zh) * 2011-12-02 2013-06-05 中国移动通信集团公司 一种推荐项的推送方法、系统及推荐服务器
CN103428169A (zh) * 2012-05-17 2013-12-04 腾讯科技(深圳)有限公司 Sns社区中推荐用户的方法和系统
CN103678518A (zh) * 2013-11-28 2014-03-26 北京邮电大学 一种推荐列表调整方法和装置
CN104615609A (zh) * 2014-04-30 2015-05-13 腾讯科技(深圳)有限公司 一种联系人管理方法及装置
CN104202319A (zh) * 2014-08-28 2014-12-10 北京淘友天下科技发展有限公司 一种社交关系推荐方法及装置

Also Published As

Publication number Publication date
CN108038789A (zh) 2018-05-15

Similar Documents

Publication Publication Date Title
WO2019085120A1 (zh) 协同过滤推荐方法、电子设备及计算机可读存储介质
US10754985B2 (en) Anonymizing user identifiable information
Korolev et al. An improvement of the Berry–Esseen inequality with applications to Poisson and mixed Poisson random sums
US9158801B2 (en) Indexing based on object type
US9661100B2 (en) Podcasts in personalized content streams
US9304979B2 (en) Authorized syndicated descriptions of linked web content displayed with links in user-generated content
US9721015B2 (en) Providing a query results page
CN110223186B (zh) 用户相似度确定方法以及信息推荐方法
WO2019062021A1 (zh) 在应用程序中推送贷款广告的方法、电子装置及介质
TW201843622A (zh) 基於資料源的業務客製裝置、方法及電腦可讀儲存介質
CN105378726A (zh) 利用更新层的数据库分区
US8959390B2 (en) Directory-level RAID
CN107408115B (zh) web站点过滤器、控制对内容的访问的方法和介质
WO2019085463A1 (zh) 部门需求的推荐方法、应用服务器及计算机可读存储介质
US20130198240A1 (en) Social Network Analysis
US10268713B2 (en) Federated social media analysis system and method thereof
WO2015081848A1 (zh) 社交化扩展搜索方法及相应的装置、系统
WO2019061664A1 (zh) 电子装置、基于用户上网数据的产品推荐方法及存储介质
US9602607B2 (en) Query-driven virtual social network group
US10049128B1 (en) Outlier detection in databases
US11074276B2 (en) Methods and systems for optimized visual summarization for sequences of temporal event data
CN111191454A (zh) 一种实体匹配的方法及装置
CN111159183B (zh) 报表生成方法、电子装置及计算机可读存储介质
WO2019041525A1 (zh) 同名实体识别方法、电子设备及计算机可读存储介质
WO2019085333A1 (zh) 项目推荐方法、电子设备及计算机可读存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18872488

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 23/09/2020)

122 Ep: pct application non-entry in european phase

Ref document number: 18872488

Country of ref document: EP

Kind code of ref document: A1