Modeling new ranking behavior of users in online social networks to achieve efficient recommending algorithms
الموضوعات : International Journal of Data Envelopment Analysis
Mehdi Safarpour
1
,
S.Hadi Yaghoubyan
2
,
Karamolah BagheriFard
3
,
razieh malekhoseini
4
,
Samad Nejatian
5
1 - Islamic Azad University Shiraz Branch
2 - Department of Computer Engineering, Yasuj Branch, Islamic Azad University, Yasuj, Iran
3 - Department of Computer Engineering, Yasooj Branch, Islamic Azad University, Yasooj, Iran
4 - Assistant professor in architecture, Department of architecture and urban design, Shiraz Branch , Islamic Azad University, Shiraz , Iran
5 - Department of Electrical Engineering, Yasooj Branch, Islamic Azad University, Yasooj, Iran
الکلمات المفتاحية: social networks, recommendation systems, collaborative filtering, users' behavior prediction, social relationship,
ملخص المقالة :
Today, the role of recommender algorithms on online platforms of social networks is considered with the penetration rate increase on the platforms. Moreover, investigating these algorithms' functionality accuracy is important in the appropriate recommendations. These algorithms have been introduced for a long time, and they try to propose appropriate recommendations by the users' behavior modeling. But, nowadays, with an increasingnumber of people relatives on social networks, the networks' users' behavior is psychological. Moreover, the people's action is related to different events such as publishing a post in addition to the post's contention, the user interest in the post publisher, and the people relationship. This paper shows that user behavior in a social network is predictable and indicates the possibility of incorrect recommendations in participatory filtering algorithms.
