查看文章

The popularity among social network helps to share their views on anything by rating, generating blogs and analysis of user behavior (review). The new type of recommendation system in movie will provide best opportunity for users to solve cold start problems. This system is based on probabilistic matrix factorization. Four major factors are used for construction of this system datasets, prediction of rating, cosine similarity and NB tree. Moreover the ratings are made by experts which will be determined by number of accurate rating that they are made on movies. The accurate results are classifying by using cosine similarity; this system combines both user interest and social circle. Cosine similarity is mainly used to judge the similarity measure between the two users; interest is compared with high rate of accuracy this helps to find the interest of user more accurately. The NB tree is used to generate the user link formation. The user curiosity was classified into various level of ranges based on the rating of movie given by review experts. Experimental results show the proposed approach outperforms the existing Recommendation System approaches.