M.Kamala Kumari

Work place: Department of CSE, AdikaviNannaya University, Rajahmundry, Andhra Pradesh, India

E-mail: kmarepalli@yahoo.com


Research Interests: Network Architecture, Network Security, Data Structures and Algorithms, Analysis of Algorithms, Mathematical Analysis


Dr. M. Kamala Kumari is presently working as Associate Professor in the Department of Computer Science and Engineering, Adikavi Nannaya University, Rajahmundry, India. She published several research papers in national and international journals and presented papers at various conferences, seminars and workshops.Her current research interests are Data Analytics, Social Network Analysis, Data Mining and Cloud Computing.

Author Articles
A Community Based Reliable Trusted Framework for Collaborative Filtering

By Satya Keerthi Gorripati M.Kamala Kumari Anupama Angadi

DOI: https://doi.org/10.5815/ijisa.2019.02.07, Pub. Date: 8 Feb. 2019

Recommender Systems are a primary component of online service providers, formulating plenty of information produced by users’ histories (e.g., their procurements, ratings of products, activities, browsing patterns). Recommendation algorithms use this historical information and their contextual data to offer a list of likely items for each user. Traditional recommender algorithms are built on the similarity between items or users.(e.g., a user may purchase the identical items as his nearest user). In the process of reducing limitations of traditional approaches and to improve the quality of recommender systems, a reliability based community method is introduced.This method comprises of three steps: The first step identifies the trusted relations of the current user by allowing trust propagation in the trust network. In next step, the ratings of selected trusted neighborhood are used for predicting the unrated item of current user. The prediction relies only on items that belong to candidate items’ community. Finally the reliability metric is computed to assess the worth of prediction rating. Experimental results confirmed that the proposed framework attained higher accuracy matched to state-of-the-art recommender system approaches.

[...] Read more.
Other Articles