Refine your search
Collections
Co-Authors
Year
A B C D E F G H I J K L M N O P Q R S T U V W X Y Z All
Sharma, Richa
- MOVBOK: A Personalized Social Network Based Cross Domain Recommender System
Abstract Views :140 |
PDF Views:0
Authors
Affiliations
1 Department of Computer Science and Engineering, Chandigarh University, IN
1 Department of Computer Science and Engineering, Chandigarh University, IN
Source
Indian Journal of Science and Technology, Vol 9, No 31 (2016), Pagination:Abstract
Objective: We propose a novel idea for resolving research issues like cross domain recommendations and recommendations using social networks in the emerging research field recommender systems. Methods/Analysis: According to this idea user will be recommended with the list of books that belong to the genre that is most liked by the user in terms of movies. Findings: Here we will collect user's tastes in movies from his social network profile and extract out the most liked genre by him and using an appropriate collaborative filtering algorithm will recommend him with the books that may interest him. Improvement: The proposed idea is expected to resolve research problems like cold start problem and sparsity. Our proposed methodology gives more competent results than the traditional.Keywords
Cross Domain Recommendations, F1 Score, Precision, Recall, Recommender Systems, Root Mean Square Error.- Community Overlapping Detection in Complex Networks
Abstract Views :167 |
PDF Views:0
Authors
Affiliations
1 Lovely Professional University, Phagwara - 144411, Punjab, IN
1 Lovely Professional University, Phagwara - 144411, Punjab, IN
Source
Indian Journal of Science and Technology, Vol 9, No 28 (2016), Pagination:Abstract
Background/Objectives: The community overlapping is the process by which number of nodes within the mesh network share common resources. The shared resources could lead to the conflict such as inconsistent analysis problem. Study of these problems is the objective of the paper. Methods/Statistical Analysis: In order to analyze the problem Modified K-Clique with sink node elimination technique is suggested. K-Clique method used detects the nodes in the mesh network having more than one connection. The modification to K-Clique enhance speed since sink node is eliminated prior to calculation of cliques. The adjacency matrix is used in order to detect the sink nodes. The Simulation is conducted in MATLAB. The MATLAB provides tools of network programming in terms of plots and graphs. The existing K-Clique is compared against the modified K-Clique and result obtained is better for Modified k-Clique. Findings: The speed is enhanced almost by 5% and number of cliques of distinct sizes discovered is also increased by 5%. The speed can further be enhanced by following hop count mechanism to reach destination quickly in addition to sink node elimination. Application/Improvement: Enhancement of performance using community overlapping detection in wireless mesh network through which it is possible to transfer the data towards multiple destinations with the help of community overlapping detection. Multiple destination towards which is to be transferred can be detected. Time will be less consumed in this case. The distance vector routing can be merged in the supposed system to further enhance the scope of the system.Keywords
Community Overlapping, Complex Network, Distance Vector, K-Clique, Sink Nodes.- Evolution of Recommender Systems from Ancient Times to Modern Era: A Survey
Abstract Views :198 |
PDF Views:0
Authors
Richa Sharma
1,
Rahul Singh
1
Affiliations
1 University Institute of Engineering, Chandigarh University, Gharuan, Mohali - 140413, Punjab, IN
1 University Institute of Engineering, Chandigarh University, Gharuan, Mohali - 140413, Punjab, IN