Structural Balance Theory-based E-commerce Recommendation over Big Rating Data (Sep-2018)
The main aim of this project is provide an user trust based and individual attribute based recommendation.
In our proposed model we can provide a user trust based recommendation and meet the user individual attribute based recommendation. In our implementation we can create two separate web application one for our own social network. The usage of social network to collect our friends and friends of list for providing our trust based recommendation. In that social network user can create account and adding the friends and chatting with particular friends, sharing the post with access control.
Next web application is our Recommendation model. In that model first we can create and upload our dataset to Hadoop file system. We can create the movie based recommendation system. Next user can enter into our application and view the list of movies and view the movies overall rating and individual rating and apply the rating to the particular movie.
Next user request the movie based on our requirements then we can collect the user requirement and apply for the collaborative filtering algorithm for three levels.
The first level filtering is collect the user friends in the social networks and find the users enemy in the friends list based on our interest. Next we can find the enemy’s friends in social network based on enemy’s interest and find the enemy’s of enemy (EOE). That EOE interest will match on request user interest.
Next apply for second level filtering, collect the EOE movie rating based on requested user requirements. That list having the top order rating as well as lower order rating.
Final level filtering is collect the top order rating movies list and provide the recommendation to the requested user and the requested user’s friend’s may be having the similar interest, so we can find that friends rating and provide the recommendation.