Recommender System
A hybrid recommender system for the movies ranking, where a movie based recommender system suggests the user about the movie that he should rank after performing the intelligent analysis using weighted approach.
Abstract
-One of the major data mining applications is Recommender System. It is the intelligent system that basically investigate the dataset present in existing system and based on which it will give some suggestions to the user regarding further process. These recommender systems are generally application specific and work on certain parameters. In this present work we define a hybrid recommender system for the movies ranking. A movie based recommender system suggests the user about the movie that he should rank after performing the intelligent analysis. In this present work, we are defining three dimensions to get the concept of hybridization. This kind of dataset having two main dimensions called users, movies and the relationship. The first level analysis will be based on user side where the content based weighted similarity analysis will be performed. Once the similar users will be identified, the next work is performed on movie side. The similar movies based on different aspects are identified using content based weighted analysis. At the third level, the similarity analysis between the relationships is identified using collaborative analysis. To perform the collaborative analysis, correlation coefficient will be used. Once these three level analysis will be completed, the next work is to conclude the relationship using weighted approach. The weightage will be applied all three methods and obtain the analysis. Another improvement here defined is the analysis under the temporal vector. It means, instead of analysis on whole dataset, the dataset in same time domain will be considered only. The work will be implemented in Matlab environment.