User Based Collaborative Fltering with Recursive Neural Network Prediction Model
S. Prasanna Priya1, M. Karthikeyan2

1S. Prasanna Priya*, Assistant Professor, Thiru A. Govindasamy Govt Arts College, Tindivanam, Tamil Nadu, India.
2Dr. M. Karthikeyan, Assistant Professor, Department of Computer and Information Science, Faculty of Science, Annamalai University, Annamalainagar, Tamil Nadu, India.
Manuscript received on February 10, 2020. | Revised Manuscript received on February 22, 2020. | Manuscript published on March 10, 2020. | PP: 1713-1718 | Volume-9 Issue-5, March 2020. | Retrieval Number: E3108039520/2020©BEIESP | DOI: 10.35940/ijitee.E3108.039520
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (

Abstract: Nowadays, with the enormous volume of online data, more consideration has been given to develop information driven recommender systems (RSs). Those schemes automatically guide consumers to discover services or movie with regard to their own interests from a huge set of possible choices. Most RS are employed to recommend the user based on their ratings and their preferences. Hence the existing RS provides very narrow recommendations and it restrict the user from accessing the different products. In this paper a novel Movie Recommender System with Cosine Similarity based Collaborative filtering and Recursive neural network MRS-CCR is proposed to the users based on the movie ratings. In the proposed RS the cosine similarity is utilized for determining the similarity among the users over the rated movies, which is employed to predict the rating of the unrated movie for each user through collaborative filtering. The Collaborative filtering (CF) is more successful recommendation methods because of its simplicity and accuracy. In the present work, matrix factorization technique is used for collaborative filtering. The obtained outcome of collaborative filter is fed into the Recursive neural network which is based on tanh activation function. The Recursive neural network predicts the recommended movies to the user. The outcome of the Recursive neural network is used for constructing the confusion matrix for evaluation. The experimental outcome of MRS-CCR is related to existing system on error and accuracy metrics. The proposed MRS-CCR has the accuracy of 95.53% better than the existing RS.
Keywords: Movie, Recommender Systems, Recursive Neural Network, Collaborative Filtering, Cosine Similarity.
Scope of the Article: Neural Information Processing