An Improved B-hill Climbing Optimization Technique for Solving the Text Documents Clustering Problem.

Area of Science:

• Computer Science
• Information Retrieval

Background:

• The internet's vast information requires efficient methods for managing large text document collections.
• Text clustering is a key optimization technique for organizing information into coherent groups.

Purpose of the Study:

• To introduce a novel local clustering technique, β-hill climbing, for text document clustering.
• To enhance the partitioning of similar documents into clusters.

Main Methods:

• The proposed β-hill climbing technique models document similarity for effective clustering.
• A key innovation is the β parameter, balancing local and global search capabilities.

Main Results:

• Experiments on eight benchmark datasets demonstrated the effectiveness of β-hill climbing.
• The proposed method outperformed the original hill climbing technique in text clustering tasks.

Conclusions:

• The addition of the β operator significantly enhances the performance of hill climbing for text clustering.
• β-hill climbing offers an improved approach to managing and organizing large volumes of text data.