1. Movie Recommendation System Addressing Changes in User Preferences with Time
- Author
-
Prakhar Yadav, Dilpreet Kaur, Kajal Parikh, Hemant Rathore, and Dhairya Parikh
- Subjects
Information retrieval ,Artificial neural network ,Mean squared error ,Computer science ,business.industry ,Order (business) ,Deep learning ,Collaborative filtering ,Hit rate ,Artificial intelligence ,Recommender system ,business ,Preference - Abstract
With a rapid increase in the volume of data, it is extremely critical for a business to display users with personalized recommendations. Applications such as Youtube, Amazon Prime Video, etc., keep track of the user’s activity, preferences, and ratings in order to provide meaningful recommendations. Such recommenders are typically based upon long-term preference profiles. However, our proposed approach implements a deep learning model that overcomes the current limitations of content and collaborative based filtering approaches and achieved a root mean squared error of 0.8659. These intermediate recommendations are further filtered based upon the changes in user preferences over time and provides the final recommendations tailored to user’s recent interactions. The proposed approach obtains a hit rate of 63% taking the user’s current interests and behavior into account.
- Published
- 2021
- Full Text
- View/download PDF