User preference modeling for movie recommendations based on deep learning

Area of Science:

• Artificial Intelligence
• Computer Science
• Information Retrieval

Background:

• Traditional movie recommendation systems struggle with complex user preferences.
• Existing methods like content-based and collaborative filtering have limitations.
• Personalized movie recommendations require advanced techniques to capture user dynamics.

Purpose of the Study:

• To introduce a novel deep learning-powered method for enhancing movie recommendation models.
• To address the limitations of current recommendation systems in capturing user preferences.
• To achieve a greater degree of customization in movie recommendations.

Main Methods:

• Utilizing Artificial Intelligence (AI), graph-based techniques, and text mining.
• Employing PageRank to rank movies based on user browsing history importance.
• Using Convolutional Neural Network (CNN) to predict user acceptance of movies.

Main Results:

• The novel approach demonstrated significant enhancements in recommendation precision and recall.
• Achieved a 7.15% increase in precision and a 5.19% increase in recall.
• Experimental evaluation conducted on a dataset of 215 users and 508 movie pages.

Conclusions:

• The proposed AI-driven method offers a substantial improvement over existing movie recommendation systems.
• The technique effectively analyzes user behavior and movie content for highly personalized suggestions.
• The findings indicate strong potential for implementing this approach in real-world personalized movie recommendation platforms.