Modelling the Behavior Classification of Social News Aggregations Users

Abstract

This paper deals with actual fuzzy logic approach for modelling the behavior classification of social news aggregations users. The peculiarities of the structure of informational content of communities on the basis of social news aggregations are explored. A formal model of social news aggregation model has been developed, which includes user of the social news aggregation on the basis of fuzzy measures of its characteristics. The method of behavioral classification of users and methods for structuring sections and discussions of social news aggregations are developed. The methods for determining the main characteristics of the users of the social news aggregation: activeness, creativeness, attractiveness, reactiveness, loyalty, is developed. Method for defining characteristics and classification of social news aggregations users is presented.

Keywords

• classification
• social network
• modelling
• social news aggregation
• fuzzy logic
• behavior classification
• web
• user

-

Each user of social news aggregations makes a contribution to the development of a social news aggregation. Contribution of the user can be determined objectively (based on the study of its behavior, information content, which is created and classification) and subjective (based on the assessments of other user of social news aggregations and expert evaluations). The indicator of user usefulness allows ranking the users of social news aggregations, using received information for further administrative measures. For example, the most useful users need to be involved in moderation of the social news aggregation, to stimulate the material, while users with negative usefulness need to remove from social news aggregations.

Ranking the users in the contribution and determining the core of social news aggregation user allows the administrator at any time to establish the users, who bring the social news aggregation the maximum benefit.

This information is necessary and critically useful for managing the social news aggregation, since the social news aggregation is different from the usual site, which the administrator must take into account the views and interests of the users. And in order to make social news aggregation-based solutions is needed to listen to users, who have a great authority and make the main contribution to its development. The development of methods for behavioral classification of users of social news aggregation based on the presentation of information content as a tree is an urgent task. 2

The Method of User Behavioral Classification The user of the social news aggregation a person who visits the social news aggregation site, reads or publishes its content in the form of discussions and messages on the social news aggregation is considered.

The model of the user of the social news aggregation is assigned in the following form:

SNAUi 

Logini ,Passwordi ,Statusi ,Emaili ,

MemNamei ,LastVisiti ,PersonalDatai

where SNAUi is a user of social news aggregations; Logini is pseudo of user; Passwordi is password of user; Statusi is the role of a user in the community; Emaili is email; Memnamei is the name of the user; LastVisiti is date of the last visit to social news aggregation; PersonalDatai is personal data of the user. 2.1

Development of Methods for the Calculating Characteristics of Social News Aggregation Users In the course of the research it was established that the users of the social news aggregations ( 1 ) to a greater or lesser extent have the following characteristics in Figure 1.

Based on these characteristics, we define the rules for classifying users in social news aggregations.

For presentation the characteristics of the users in the social news aggregation we will use a set of unclear plurals of the listed characteristics of the community users in relation to the entire social news aggregation are determined based on the analysis of the behavior of users within the community: ─ activeness is determined by the amount of information content they create; ─ creativeness is determined the quality of information content and how other users of community react to it; ─ attractiveness is determined the quantity of users who react to the created content; ─ reactiveness is a way of participating in discussions; ─ loyalty is a reaction to the information content of other users. 2.2

Development of

Methods for

Calculating the

Values of Linguistic Variables and Measures of Belonging

For each of the proposed characteristics of the users, we introduce the corresponding linguistic variables: Activeness, Creativeness, Attractiveness, Reactiveness, and Loyalty.

The linguistic variable is given by the quartet < , , , > where is the name of the linguistic variable; T is a plural of values of a linguistic variable, which is the names of fuzzy variables; T  " low"," medium"," high" = [0; ( )]; X is area of definition of fuzzy variables describing the linguistic variable M is a set of measures of fuzzy variables, which are values of a linguistic variable. 1  0,5

A

2 A 1

A 3 x Pни1зька

Pсе1редня

Pни2зька

Pсе2редня

Pсе3редня

Pви1сока

Pсе4редня

Pви2сока card(Post) eters that are proportional to card(Post) and given by an expert or administrator of the forum, and moreover Pl1ow  Pm1edium  Plo2w  Pme2dium  Pm3edium  Phi1gh  Pme4dium  Phi2gh  1 . All the P parameters that are used later in the membership functions are determined by experts for each feature of the social news aggregation.

We will write down the functions of membership for all characteristics ( 1 ) – (30) of the users of the social news aggregation.

The activeness of creating discussions: 

Th   1,if Pat m2edium  ActivenessThread  SNAU i   Pat m3edium where ActivenessThread  SNAU  is activeness of creating posts j-th user.

i ( 1 ) ( 2 ) ( 3 )

Patmedium  ActivenessThread  SNAUi  ,

3 3

Patmedium  Patmedium  if Pat m3edium  ActivenessThread  SNAUi   Pat m4edium   high  Activeness    Th   if Pat 1 high    2 1, if Pathigh

Thread Pat 2 high  SNAU   Pat i 1 high  Pat 1 high

,  Activeness

Thread  SNAU   Pat i 2 high  Activeness

Thread  SNAU   1 i 

low

high 

low where ActivenessPoll  SNAUi  is the activity of creating polls i-th user.

The activeness of creating posts:

Paplh1igh  ActivenessPoll  SNAUi   Paplhi2gh

Paplhi2gh  ActivenessPoll  SNAUi   1 , , 1 low  ActivenessPost 2  SNAU i   Papslow Papl

Papsm4edium  ActivenessPost  SNAU i  ,

low  1,   Pvt Vt    if

high where posts.  high

Pvth1igh  ActivenessVote  SNAUi   Pvthi2gh 1  ActivenessVote  SNAUi   Pvthigh  2 1  Pvthigh  Pvthigh  Vt   if   

1, if 

Pvt h2igh  ActivenessVote  SNAUi   1 where ActivenessVote  SNAU  is active participation in the voting.

i

The activeness of evaluating the actions of other users:  low  1, if 0  ActivityFeedback   Pfblo2w  ActivityFeedback  Fb    Pfblo2w  Pfblo1w  1 if Pfblow  ActivityFeedback  SNAU   Pfb i ,

Pfbhi2gh  ActivenessFeedback  SNAUi   1 Phi1gh  ActivenessFeedback  meSNAUi mberi   Pfbhi2gh ( 12 ) ( 13 ) ( 14 ) ( 15 ) where ActivenessFeedback  SNAU  is the activeness of evaluating actions by the ii th user.

Total activeness:

Calculated based on these types of activeness:  low 1,   Total       якщо 0  ActivenessTotal 1  SNAU   Ptlow i Ptlo2w  ActivenessTotal  SNAU i  ,

2 1

Ptlow  Ptlow if

Ptlo1w  ActivenessTotal  SNAUi   Ptlo2w  ActivenessTotal  SNAUi   1 1 if 0  Creativeness  SNAU i   Pcrlow 2  Creativeness  SNAU  low i

Pcr lo1w  Creativeness  SNAU i   Pcrlo2w 

high  low Creativeness of the user: 1,   Pcr Cr     

if Pcr  ( 16 ) ( 17 ) ( 18 ) ( 19 ) 1  ActivenessTotal  SNAUi   Pthigh  Pt h2igh  Pth1igh  1 Total   if Pthigh    1, if Pt h2igh where ActivenessTotal  SNAU  is total activeness of the i-th user.

i  if Pat m3edium  ActivenessTotal  SNAUi   Pat m4edium  , ,        medium Cr   1,

Creativeness  SNAUi   Pcrm1edium ,

Pcrm2edium  Pcrm1edium if Pcrm1edium  Creativeness  SNAUi   Pcrm2edium

if Pcrm2edium  Creativeness  SNAUi   Pcrm3edium Pcrm4edium  Creativeness  SNAU 

i ,

Pcrm3edium  Pcrm3edium

Pcrm3edium  Creativeness  SNAUi   Pcrm4edium  1, if 0  Attractiveness  SNAUi   Pattrlo1w   Pattrlo2w  Attractiveness  SNAU    Attr   1 i , low

Pattrlo2w  Pattrlow

1    if Pattrlo1w  Attractiveness  SNAUi   Pattrlo2w  Attractiveness  SNAUi   Pattrm1edium ,  2 1  Pattrmedium  Pattrmedium  1  if Pattrmedium  Attractiveness SNAUi   Pattrm2edium   medium  Attr   11, if Pattrm2edium  Attractiveness SNAUi   Pattrm3edium Creativeness  SNAU   Pcrhi1gh  i ,  Pcrhi2gh  Pcrhi1gh   high Cr   if Pcrhi1gh  Creativeness  SNAUi   Pcrhi2gh     1, if Pcrhi2gh  Creativeness  SNAU   1 i where Creativeness  SNAU  is creativeness of the user.

i Аtractiveness of the user:

4 Pattrmedium  Attractiveness SNAU 

i , 3 3

Pattrmedium  Pattrmedium if Pattrmedium  Attractiveness SNAUi   Pattrm4edium

3 where Attractiveness  SNAUi  is attractiveness of the user.

Reactiveness of the user.:  1, if 0  Reactiveness SNAU   Prlo1w

i   Prlo2w  Reactiveness  SNAU    R  1 i , low

Prlo2w  Prlo1w    if Prlo1w  Reactiveness  SNAUi   Prlo2w  medium  R  1, if Prm2edium  Reactiveness SNAUi   Prm3edium  Reactiveness SNAUi   Prm1edium ,   Prm2edium  Prm1edium   if Prm1edium  Reactiveness SNAUi   Prm2edium     Prm4edium  Reactiveness SNAUi  ,

Prm3edium  Prm3edium  if Prm3edium  Reactiveness SNAUi   Prm4edium   Reactiveness  SNAUi   Prhi1gh ,   Prhi2gh  Prhi1gh  high  R  if Prhi1gh  Reactiveness  SNAUi   Prhi2gh     1, if Prhi2gh  Reactiveness  SNAUi   1 where R e activeness SNAUi  is reactiveness of the user.

Loyalty of the user: (25) (26) (27) 1, if 0  Loyalty  SNAUi   Pllo1w  Pllo2w  Loyalty  SNAU   low  L    i ,  Pllo2w  Pllo1w   if Pllo1w  Loyalty  SNAUi   Pllo2w  Loyalty  SNAU i   Pl m1edium ,   Pl m2edium  Pl m1edium   if Pl m1edium  Loyalty  SNAU i   Pl m2edium   medium  L   1, if Pl m2edium  Loyalty  SNAU i   Pl m3edium   Pl m4edium  Loyalty  SNAU i  ,  Pl m3edium  Pl m3edium  if Pl m3edium  Loyalty  SNAU i   Pl m4edium  Loyalty  SNAUi   Plh1igh ,   Plhi2gh  Plh1igh   high  L   if Plh1igh  Loyalty  SNAUi   Plhi2gh     1, if Plhi2gh  Loyalty  SNAUi   1 where Loyalty  SNAUi  is loyalty of the user. (29) (30) 3

Building Rules Aggregation for

Classifying

Users of

Social

News The classification rules for each class of users of the social news aggregation are formulated based on the developed methods for calculating the characteristics of users and certain classes of social news aggregation users. The classes of the social news aggregation are proposed:  Activist  Moderator  Flamer  Author  Critic  Reader.

The membership of the users in one of the classes based on its characteristics (Activity, Creativity, Attraction, Reactivity, Loyalty) is represented by production rules and Table 1.

If Асtiveness(SNAU)"medium","high" and Creativeness(SNAU)"medium","high" and Reactivenes(SNAU)"low","medium" then Member - Аctivist; If Аctiveness(SNAU)"medium","high" and Reactivenes(SNAU) ="high" and Loyatly(SNAU)"medium","high" , then Member - Мoderator; If Аctiveness(SNAU) "medium","high" and Creativeness(SNAU) = "low" and Loyatly(SNAU) = "low", then Member - Babbler; If Аctiveness  SNAU  = "low" and Creativeness(SNAU)"medium","high" and Аtractiveness(SNAU)"medium","high" and Reactiveness  SNAU  = "low",then Member - Аuthor; If Creativeness  SNAU  = "low" and Rеаctiveness (SNAU) "low","high" and Аtractiveness(SNAU) = "high" and Loyatly  SNAU  = "low",then Member - Critic; If Аtractiveness  SNAU  = "low" and Creativeness( SNAU ) = "low" and Atractiveness( SNAU )"low","average" and

Reactiveness  SNAU  = "low", then Member - Reader;

Determine the User's Usefulness for the Community The usefulness of a user for the social news aggregation is a complex indicator calculated on the basis of its characteristics: activeness, attractiveness, creativeness, reactiveness and loyalty. The usefulness of user ME is calculated by the equation: over Ci  1 , Ci  0 . determined based on the development scenario of the social news aggregation, morei Consequently, ∈ [0, 1].

The user's usefulness allows the administrator to evaluate the importance of the user for the community and to take this value into account when applying sanctions.

is weight coefficients of each user's characteristics, which are 5

Conclusion In this work the models have been developed that are the basis for further research on the construction of effective site positioning methods. Formalized structure of the social news aggregation, which includes two components (information content, users) is suggested.

The peculiarities of the structure of informational content of communities on the basis of social news aggregations are explored. A formal social news aggregation model has been developed, which includes the model of user of the social news aggregation on the basis of fuzzy measures of its characteristics, the model of the structure of information content and the model of the content of information content, on the basis of which developed the method of behavioral classification of users and methods for structuring sections and discussions of social news aggregations.

The methods for determining the main characteristics of the users of the web community: activeness, creativeness, attractiveness, reactiveness, loyalty are developed. The classes of users of social news aggregations on the basis of social news aggregations are allocated and the rules of classification of users are formulated.

References

1. Jin , L. , Chen , Y. , Wang , T. , Hui , P. , Vasilakos , A. : Understanding user behavior in online social networks: a survey . Communications Magazine , 51 , ( 9 ), 144 - 150 ( 2013 ).

2. Benevenuto , F. , Rodrigues , T. , Cha , M. , Almeida , V. : Characterizing user behavior in online social networks . In: 9th ACM SIGCOMM conference on Internet measurement (IMC '09) . ACM, pp. 49 - 62 . New York, USA ( 2009 ).

3. Fernández , A. , García , S. , del Jesus , M. J. , Herrera , F. : A study of the behaviour of linguistic fuzzy rule based classification systems in the framework of imbalanced data-sets. Fuzzy Sets Syst ., 159 ( 18 ), 2378 - 2398 ( 2008 ).

4. Korzh , R. , Peleshchyshyn , A. , Syerov , Y. , Fedushko , S.: University' s information image as a result of university web communities' activities . Advances in Intelligent Systems and Computing. Advances in Intelligent Systems and Computing , 512 , pp. 115 - 127 . Springer, Cham ( 2017 ). DOI: 10 .1007/978-3- 319 -45991- 2 _ 8

5. Morente-Molinera , J. , et al.: Supervised learning classification methods using multigranular linguistic modeling and fuzzy entropy , Transactions on Fuzzy Systems , 25 ( 5 ), 1078 - 1089 ( 2017 ).

6. Fedushko , S. , Ustyianovych , T. : Predicting pupil's successfulness factors using machine learning algorithms and mathematical modelling methods. Advances in Computer Science for Engineering and Education II . ICCSEEA 2019. Advances in Intelligent Systems and Computing , 938 , pp. 625 - 636 . Springer ( 2020 ). DOI 10 .1007/978-3- 030 -16621-2_ 58

7. Olson , D. L. , Delen , D: Advanced Data Mining Techniques . NY, USA: Springer, 2008 .

8. Lee , H. -M. , et al.: An efficient fuzzy classifier with feature selection based on fuzzy entropy , Trans. Syst. Man Cybern. B Cybern ., 31 ( 3 ), 426 - 432 ( 2001 ).

9. Wendy , N. , Misha , P. : Moving behavioral theories into the 21st century: technological advancements for improving quality of life . IEEE pulse. 4 ( 5 ), 25 - 28 ( 2013 ).

10. Syerov , Y. , Fedushko , S. , Loboda , Z. : Determination of development scenarios of the educational web forum . In: 11th International Scientific and Technical Conference “Computer Sciences and Information Technologies” CSIT-2016 , 73 - 76 ( 2016 ). DOI: 10 .1109/STCCSIT. 2016 . 7589872 .

11. Liu , J. , Weitzman , E. , Chunara , R. : Assessing Behavioral Stages From Social Media Data . In: Conference on Computer-Supported Cooperative Work, pp. 1320 - 1333 ( 2017 ).

12. Bilushchak , Т., Peleshchyshyn , A. , Komova , M. : Development of method of search and identification of historical information in the social environment of the Internet . In: XIth International Scientific and Technical Conference "Computer Sciences and Information Technologies" CSIT-2017 , 196 - 199 ( 2017 ).

13. Trach O. , Peleshchyshyn A. : Functional-network model of tasks performance of virtual communication life cycle directions . In: XIIth International Conference "Computer Sciences and Information Technologies" CSIT-2016 . pp. 108 - 110 . Lviv ( 2016 ).

14. Trach O. , Vus V. , Tymovchak-Maksymets O. : Typical algorithm of stage completion when creating a virtual community of a HEI . In: XIIIth International Conference on Advanced Trends in Radioelectronics, Telecommunications and Computer Engineering, TCSET-2016, pp. 849 - 851 . Lviv-Slavske ( 2016 ).

15. Dosyn , D. , Lytvyn , V. , Kovalevych , V. , Oborska , O. , Holoshchuk , R. : Knowledge discovery as planning development in knowledgebase framework. Modern Problems of Radio Engineering, Telecommunications and Computer Science . In: 13th International Conference on on Advanced Trends in Radioelectronics , Telecommunications and Computer Engineering, TCSET-2016, pp. 449 - 451 . Lviv-Slavsko, Ukraine ( 2016 ).

16. Lytvyn , V. , Peleshchak , I. , Peleshchak , R. , Holoshchuk , R.: Detection of multispectral input images using nonlinear artificial neural networks . In: 14th International Conference on Advanced Trends in Radioelectronics , Telecommunications and Computer Engineering, TCSET, pp. 119 - 122 . Slavske, Ukraine ( 2018 ).

17. Artem , K. , Holoshchuk , R. , Kunanets , N. , Shestakevysh , T. , Rzheuskyi , A. : Information Support of Scientific Researches of Virtual Communities on the Platform of Cloud Services . Advances in Intelligent Systems and Computing. Advances in intelligent systems and computing III , 871 , pp. 301 - 311 ( 2019 ).

18. Zhezhnych , P. , Tarasov , D. : Methods of data processing restriction in ERP Systems . In: 13th International Scientific and Technical Conference on Computer Sciences and Information Technologies , CSIT 2018 , pp. 274 - 277 ( 2018 ).

19. Tkachenko , R. , Izonin , I. : Model and Principles for the Implementation of Neural-Like Structures based on Geometric Data Transformations. Advances in Computer Science for Engineering and Education . International Conference on Computer Science, Engineering and Education Applications, ICCSEEA 2018 . Advances in Intelligent Systems and Computing , vol 754 , 578 - 587 . Springer, Cham ( 2019 ). doi: 10 .1007/978-3- 319 -91008-6_ 58

20. Gozhyj , A. , Vysotska , V. , Yevseyeva , I. , Kalinina , I. , Gozhyj , V. : Web resources management method based on intelligent technologies . Advances in Intelligent Systems and Computing , 871 , 206 - 221 ( 2019 ).

21. Gozhyj , A. , Kalinina , I. , Vysotska , V. , Gozhyj , V. : The method of web-resources management under conditions of uncertainty based on fuzzy logic . In: 13th International Scientific and Technical Conference "Computer Sciences and Information Technologies" CSIT 2018 , pp. 343 - 346 ( 2018 ).

22. Lytvyn , V. , Dosyn , D. , Emmerich , M. , Yevseyeva , I. : Content formation method in the web systems . CEUR Workshop Proceedings , 2136 , pp. 42 - 61 ( 2018 ).

23. Anisimova O. , Vasylenko V. : Social networks as the instrument for a higher education institution image creation . 1-st International Workshop Control, Optimization and Analytical Processing of Social Networks ( 2019 ). In press.