Page 25 - 2018

Ph.D.
(Computer Applications)
ISSUES IN PERSONALIZED
RECOMMENDATION FOR WEB

Ph.D. Scholar : Angira Amit Patel
Research Supervisor : Dr. J. N. Dharwa

Regi. No.: 11146041001
Abstract :
The development of personalized hybrid recommendation model is considered as
challenging task due to many domain specific problems and many technical limits to
work into integrated web environment. The novel idea of this research work is
development of unique hybrid personalized recommendation model using graph data
which could help to resolved cold start and sparsity issue. The presented model is
combined with unique algorithm proposed here, which get succeed to generate results by
integrating content-based, knowledge-based and preference based recommendation
techniques. The proposed model is efficient, flexible and allows integration of multiple
entities together. Along with that, this research work emphasis on study and analysis of
various recommendation techniques, various issues related to recommendation system
and various evaluation methodology related to recommendation system. The research
work also describes various properties of graph data like flexibility, associativity
operation, diversity of structure and its suitability to work with Big data. This investigation
demonstrates use of graph structure for various recommendation algorithms are more
appropriate. This research work also demonstrate implementation of various diverse
graph models intended to design for various domains that helps in solving challenges
related to real world problems. This research serve as guidelines to anyone who wants to
implement recommendation system using graph data as integrated component of real
time web environment for their computational challenges.

Keywords: Recommendation System, Hybrid Recommendation System, Graph Database,
Big Data, NoSQL, Neo4j, Property Graph.

06