A new collaborative recommendation approach based on users clustering using artificial bee colony algorithm.

Area of Science:

• Computer Science
• Artificial Intelligence
• Data Mining

Background:

• Existing collaborative filtering recommendation methods struggle to balance accuracy and diversity in meeting user preferences.
• K-means clustering is a common technique but is susceptible to local optima, limiting its effectiveness in recommendation systems.

Purpose of the Study:

• To propose a novel collaborative filtering recommendation approach that enhances both accuracy and diversity.
• To address the local optimal problem inherent in K-means clustering for recommendation tasks.

Main Methods:

• Utilized K-means clustering algorithm for user grouping.
• Integrated the artificial bee colony (ABC) algorithm to optimize K-means clustering and avoid local optima.
• Employed modified cosine similarity to calculate user similarity within clusters for refined recommendations.

Main Results:

• The proposed method demonstrated superior performance compared to existing recommendation techniques.
• Numerical analysis on MovieLens and a real-world dataset confirmed the effectiveness of the user clustering approach.
• The integration of ABC algorithm successfully mitigated the local optimal issue in K-means clustering.

Conclusions:

• The novel collaborative filtering approach significantly improves recommendation accuracy and diversity.
• The hybrid clustering method offers a robust solution for enhancing user-based recommendation systems.
• This research provides a valuable advancement in personalized recommendation technologies.