Social Recommendation for Interactive Online System
Abstract: The research communities of information retrieval, machine learning and data mining has attracted lot of attention by a social recommendation system. Online Shopping has become a popular trend as there is no need to go hunting for products in retail markets. Social recommendation plays important role in online system where data arrives sequentially and user preference may change rapidly. To improve the linear optimization of the system using OGRPL-FW this utilizes Frank Wolfe algorithm. This Linearization induces a sublinear rate of convergence and zigzagging behavior. In this paper, we present a new framework in order to avoid these problem a regularization penalty term is added. Using collaborative filtering as well as association rules to find the missing values and predict the items efficiently. It also provides Web Based Application with user-friendly interface. The communication among the customers is provided through online chat facility, wherein they can discuss about the products for reviews. Recommendation system through web is necessary to meet the dynamically changing user preferences.
Keywords: OGRPL-FW, Regularization Penalty, Association Rules, Collaborative filtering, Chat facility
How to Cite?: Thenmalar V, Tamilselvi R, Sandhiya S, Dhivya Shree M, "Social Recommendation for Interactive Online System", Volume 6 Issue 2, February 2017, International Journal of Science and Research (IJSR), Pages: 2154-2157, https://www.ijsr.net/getabstract.php?paperid=ART20171189, DOI: https://dx.doi.org/10.21275/ART20171189