CN110851719A - Mashup Web API personalized recommendation based on collaborative filtering and link prediction - Google Patents

Mashup Web API personalized recommendation based on collaborative filtering and link prediction Download PDF

Info

Publication number
CN110851719A
CN110851719A CN201911096185.XA CN201911096185A CN110851719A CN 110851719 A CN110851719 A CN 110851719A CN 201911096185 A CN201911096185 A CN 201911096185A CN 110851719 A CN110851719 A CN 110851719A
Authority
CN
China
Prior art keywords
user
web api
web
representing
api
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
CN201911096185.XA
Other languages
Chinese (zh)
Inventor
徐九韵
孙忠顺
蔺学东
刘杰
张超
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.)
China University of Petroleum East China
Original Assignee
China University of Petroleum East China
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 China University of Petroleum East China filed Critical China University of Petroleum East China
Priority to CN201911096185.XA priority Critical patent/CN110851719A/en
Publication of CN110851719A publication Critical patent/CN110851719A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9536—Search customisation based on social or collaborative filtering
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9535—Search customisation based on user profiles and personalisation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The invention belongs to the field of recommendation, and particularly relates to recommending web API services meeting user requirements according to the requirements input by a user, and improving the accuracy and individuation of recommendation. The method mainly comprises the following four steps: mashup clustering; a user link prediction algorithm; C. a collaborative filtering algorithm based on link prediction; popularity prediction for web APIs. E.Web API recommendation model. The invention combines the link prediction and the collaborative filtering, improves the recommendation accuracy and also improves the recommendation individuation, and the recommendation results are more in line with the actual requirements of the user.

Description

Mashup Web API personalized recommendation based on collaborative filtering and link prediction
Technical Field
The invention belongs to the field of recommendation, and particularly relates to a method for recommending a web API meeting the user requirements according to the requirements of a user, so that the accuracy and individuation of a recommendation result are ensured.
Background
With the rapid development of Web services, the Web services released on the Internet are increasing day by day. In the face of a large number of Web services, especially when facing a plurality of functionally identical or similar candidate services, how to dynamically select the service that can best meet the user's needs is an important issue in the field of service discovery. Currently, as the demand of users increases, it is difficult for a single Web service to satisfy the requests of multiple functions of the users. In order to reconstruct the existing Web service resources, speed up the system construction, and reduce the cost of system development, it is necessary to combine a plurality of existing Web services according to functions, semantics, and logical relationships between them to meet the diversity of functional and non-functional requirements of users, such as cost, reputation, reliability, security, privacy, and so on. For the reasons, Mashup is used as an application development mode for quickly integrating data, and information related to a certain theme can be quickly integrated together to meet the application requirements of a contextual model. The contextual Mashup application requires that the construction can be relatively fast, and the optimal choice is achieved by utilizing an open API and a tool. The Mashup development has many advantages, such as the period of software development is accelerated, the development cost is saved, the environment is easy to build, and the integration is easy, so that common users with a certain programming basis can also carry out application development, are favored by most software development companies, and become a mainstream mode of software development.
The Web API recommendation system has the function of quickly selecting a high-quality Web API which can most meet the user requirements from massive APIs, and saving the search of the Web API in the user development process so as to quickly and efficiently develop a high-quality software service system.
Disclosure of Invention
1. The Mashup Web API personalized recommendation based on collaborative filtering and link prediction mainly comprises the following five steps:
mashup clustering: clustering the Mashup according to Mashup description information input by a user, obtaining N mashups with high similarity to the Mashup description information of the user, and obtaining Web API information contained in the N mashups.
User link prediction algorithm: the user social network is constructed according to whether the user uses the same web API and whether the same web API is commented on. Then, obtaining M similar users with higher similarity to the target user according to the similarity of the web API historically used by the similar users and the target user and the number of the web API comments, wherein the specific calculation method is as follows:
Figure BDA0002268412670000021
wherein Γ (u)1) Representing user u1Set of web APIs for comments, sim (a)i,aj) Denotes aiAnd ajN represents the number of web APIs used by the user.
C. Collaborative filtering algorithm based on link prediction: after obtaining the global similarity of the users through link prediction, the calculation method for the web API to be recommended is as follows:
Figure BDA0002268412670000022
wherein simlink(u1,u2) Representing the global similarity of users, I representing the union of the web APIs used by users u, v, ru,iRepresenting the number of times user u uses i (api), sim (u, v) representing the similarity of users.
Figure BDA0002268412670000023
Wherein
Figure BDA0002268412670000024
Representing the average rating of user u for the used web API,
Figure BDA0002268412670000025
indicating that user v scored the average of the used webAPI,
Figure BDA0002268412670000026
representing user v versus web API aiSim (u, v) represents the similarity of users u, v,representing a predicted user pair aiScoring of (4).
Web API popularity prediction:
Figure BDA0002268412670000028
wherein a isiIndicating a need to compute popularity web API, Freq (a)i) Denotes aiNumber of calls, MinValue (a)j) Denotes the minimum number of times all web APIs are called, MaxValue (a)z) Indicating the maximum number of times all web APIs are invoked,
Figure BDA0002268412670000029
denotes aiFor a further period of time with a rate of increase in the number of calls, Follow (a)i) Denotes aiNumber of followers, FollowMinValue (a)x) Represents the minimum value of followers in the web API, FollowMaxValue (a)y) Representing the maximum value of followers in the web API. Since the above data is dynamically changing, the popularity of the web API is calculated using a formula containing the above variables.
E, Web API recommendation algorithm: and adding the web API formula obtained by collaborative filtering to a corresponding popularity formula according to a proportion to obtain a final recommended value of each web API, wherein the value range of lambda is 0-1.
Figure BDA0002268412670000031
Compared with the prior art, the invention has the following remarkable advantages:
1. and constructing a social network model of the user based on whether the user uses or reviews the same API, predicting similar users of the target user according to the social network model link, and recommending the Web API which is in line with the target user through collaborative filtering.
2. And in consideration of the dynamic calling change of the Web API by the mashup, constructing a popularity model of the API and predicting the popularity of the API in the future.
Drawings
FIG. 1 is a general flow diagram of the present invention.
FIG. 2 shows recommendation results with the same Mashup requirements, different target users and different target user categories.
FIG. 3 shows recommendation results with the same Mashup requirements, the same target users and different types of target users.
FIG. 4 shows the recommendation results of Mashup with the same requirements and different target users but the same type of target users.
FIG. 5 is a comparison of the results of the model algorithm recommendation index and other recommendation algorithm indexes of the same type.
FIG. 6 is the effect of λ on the algorithm accuracy, recall and F-measure.
Detailed Description
Embodiments of the present invention will be described below with reference to the drawings.
Fig. 1 is a general flow chart of the present invention, and the recommendation algorithm based on collaborative filtering and link prediction is implemented as follows:
mashup clustering: clustering the Mashup according to Mashup description information input by a user, obtaining N mashups with high similarity to the Mashup description information of the user, and obtaining Web API information contained in the N mashups. As shown in fig. 1.
User link prediction algorithm: the user social network is constructed according to whether the user uses the same web API and whether the same web API is commented on. Then, obtaining M similar users with higher similarity to the target user according to the similarity of the web API historically used by the similar users and the target user and the number of the web API comments, wherein the specific calculation method is as follows:
Figure BDA0002268412670000041
wherein Γ (u)1) Representing user u1Set of web APIs for comments, sim (a)i,aj) Denotes aiAnd ajN represents the number of web APIs used by the user.
C. Collaborative filtering algorithm based on link prediction: after obtaining the global similarity of the users through link prediction, the calculation method for the web API to be recommended is as follows:
Figure BDA0002268412670000042
wherein simlink(u1,u2) Representing the global similarity of users, I representing the union of the web APIs used by users u, v, ru,iRepresenting the number of times user u uses i (api), sim (u, v) representing the similarity of users.
WhereinRepresenting the average rating of user u for the used web API,
Figure BDA0002268412670000045
indicating that user v scored the average of the used webAPI,
Figure BDA0002268412670000046
representing user v versus web API aiSim (u, v) represents the similarity of users u, v,
Figure BDA0002268412670000047
representing a predicted user pair aiScoring of (4).
Web API popularity prediction:
Figure BDA0002268412670000048
wherein a isiIndicating a need to compute popularity web API, Freq (a)i) Denotes aiNumber of calls, MinValue (a)j) Denotes the minimum number of times all web APIs are called, MaxValue (a)z) Indicating the maximum number of times all web APIs are invoked,
Figure BDA0002268412670000049
denotes aiFor a further period of time with a rate of increase in the number of calls, Follow (a)i) Denotes aiNumber of followers, FollowMinValue (a)x) Represents the minimum value of followers in the web API, FollowMaxValue (a)y) Representing the maximum value of followers in the web API. Since the above data is dynamically changing, the popularity of the web API is calculated using a formula containing the above variables.
E, Web API recommendation algorithm: and adding the web API formula obtained by collaborative filtering to a corresponding popularity formula according to a proportion to obtain a final recommended value of each web API, wherein the value range of lambda is 0-1.
Figure BDA0002268412670000051
According to the algorithm, when the requirements of the target users are consistent, experiments are carried out when the target users are different and the types of the target users are different, and the result of the graph 2 is obtained; obtaining the experimental result of fig. 3 when the target users are the same and the types of the target users are different; the results of fig. 4 are obtained when the target users are different and the types of the target users are the same. FIG. 5 is a comparison of this algorithm with other Web API recommendation algorithms of the same type. Fig. 6 shows the influence of a dynamic change λ on the corresponding evaluation criterion.

Claims (1)

1. The Mashup Web API personalized recommendation based on collaborative filtering and link prediction mainly comprises the following five steps:
mashup clustering: clustering mashups according to Mashup description information input by a user, acquiring N mashups with high similarity to the Mashup description information of the user, and acquiring Web API information contained in the N mashups;
user link prediction algorithm: the user social network is constructed according to whether the user uses the same web API and whether the same web API is commented on. Then, M similar users with higher similarity to the target user are obtained according to the similarity of the web API historically used by the similar users and the target user and the number of webAPI comments, and the specific calculation method is as follows:
Figure FDA0002268412660000011
wherein Γ (u)1) Representing user u1Set of web APIs for comments, sim (a)i,aj) Denotes aiAnd ajN represents the number of web APIs used by the user;
C. collaborative filtering algorithm based on link prediction: after obtaining the global similarity of the users through link prediction, the calculation method for the web API to be recommended is as follows:
Figure FDA0002268412660000012
wherein simlink(u1,u2) Representing the global similarity of users, I representing the union of the web APIs used by users u, v, ru,iRepresenting the times of using i (API) by the user u, and sim (u, v) representing the similarity of the user;
Figure FDA0002268412660000013
whereinRepresenting the average rating of user u for the used web API,
Figure FDA0002268412660000015
represents the average score of user v on a used web API,
Figure FDA0002268412660000016
representing user v versus web API aiSim (u, v) represents the similarity of users u, v,
Figure FDA0002268412660000017
representing a predicted user pair aiScoring;
web API popularity prediction:
Figure FDA0002268412660000018
wherein a isiIndicating a need to compute popularity web API, Freq (a)i) Denotes aiNumber of calls, MinValue (a)j) Denotes the minimum number of times all web APIs are called, MaxValue (a)z) Indicating the maximum number of times all web APIs are invoked,
Figure FDA0002268412660000021
denotes aiFor a further period of time with a rate of increase in the number of calls, Follow (a)i) Denotes aiNumber of followers, FollowMinValue (a)x) Represents the minimum value of followers in the web API, FollowMaxValue (a)y) Representing the maximum value of followers in the web API. Since the above data is dynamically changing, the popularity of the web API is solved using a formula containing the above variables;
e, Web API recommendation algorithm: and adding the web API values obtained by collaborative filtering to the corresponding popularity values according to a proportion to obtain the final recommended value of each web API, wherein the lambda value is 0-1.
Figure FDA0002268412660000022
CN201911096185.XA 2019-11-11 2019-11-11 Mashup Web API personalized recommendation based on collaborative filtering and link prediction Pending CN110851719A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911096185.XA CN110851719A (en) 2019-11-11 2019-11-11 Mashup Web API personalized recommendation based on collaborative filtering and link prediction

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911096185.XA CN110851719A (en) 2019-11-11 2019-11-11 Mashup Web API personalized recommendation based on collaborative filtering and link prediction

Publications (1)

Publication Number Publication Date
CN110851719A true CN110851719A (en) 2020-02-28

Family

ID=69601127

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911096185.XA Pending CN110851719A (en) 2019-11-11 2019-11-11 Mashup Web API personalized recommendation based on collaborative filtering and link prediction

Country Status (1)

Country Link
CN (1) CN110851719A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111753207A (en) * 2020-06-29 2020-10-09 华东师范大学 Collaborative filtering model of neural map based on comments

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111753207A (en) * 2020-06-29 2020-10-09 华东师范大学 Collaborative filtering model of neural map based on comments
CN111753207B (en) * 2020-06-29 2021-03-26 华东师范大学 Collaborative filtering method for neural map based on comments

Similar Documents

Publication Publication Date Title
US10726427B2 (en) Method and apparatus for building a user profile, for personalization using interaction data, and for generating, identifying, and capturing user data across interactions using unique user identification
US9535897B2 (en) Content recommendation system using a neural network language model
US20170206276A1 (en) Large Scale Recommendation Engine Based on User Tastes
KR20220041704A (en) Multi-model training method and device based on feature extraction, an electronic device, and a medium
US20160092699A1 (en) Privacy-preserving cookies for personalization without user tracking
US20210049478A1 (en) Feature relationship recommendation method, apparatus, computing device, and storage medium
Sarker et al. Mining user behavioral rules from smartphone data through association analysis
CN112836130A (en) Context-aware recommendation system and method based on federated learning
Yu et al. Time-aware collaborative filtering for QoS-based service recommendation
CN111046237B (en) User behavior data processing method and device, electronic equipment and readable medium
US10754976B2 (en) Configuring image as private within storage container
CN102170449A (en) Web service QoS prediction method based on collaborative filtering
CN108629010B (en) Web service recommendation method based on theme and service combination information
AU2020202605A1 (en) User data sharing method and device
CN115712657A (en) User demand mining method and system based on meta universe
Yao et al. Fedps: A privacy protection enhanced personalized search framework
CN111552580A (en) Distributed micro-service decision-making method, device, equipment and computer storage medium
CN110851719A (en) Mashup Web API personalized recommendation based on collaborative filtering and link prediction
Cao et al. Web API recommendation via combining graph attention representation and deep factorization machines quality prediction
US8818337B2 (en) Application of community-defined descriptors to mobile content
EP2924623A1 (en) Context based service technology
Blasch et al. Information fusion mangement and enterpise processing
CN112632569B (en) Data processing method, system, device, electronic equipment and readable storage medium
CN104954465B (en) One kind is suitable for privacy policy synthetic method under cloud service combine scenes
Wang et al. Efficient service selection in mobile information systems

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