Clustering Ensemble Based on Hybrid Multiview Clustering.

Area of Science:

• Data Science
• Machine Learning
• Artificial Intelligence

Background:

• Clustering ensemble algorithms integrate diverse clustering solutions to improve efficiency.
• Enhancing base learner diversity and optimizing ensemble strategies are key challenges.

Purpose of the Study:

• To propose a novel clustering ensemble framework to address existing challenges.
• To improve the diversity of base learners and optimize ensemble strategies.

Main Methods:

• A framework integrating three view transformation methods (random principal component analysis, random nearest neighbor, modified fuzzy extension model) as base learners.
• A random transformation and hybrid multiview learning-based clustering ensemble method (RTHMC) for synthesizing multiview results.
• Integration of random subspace transformation and a view-based self-evolutionary strategy for performance optimization.

Main Results:

• The proposed framework demonstrates effectiveness in clustering various data types.
• Experimental comparisons validate the superiority of the developed method.
• Enhanced clustering efficiency and accuracy were observed.

Conclusions:

• The proposed clustering ensemble framework offers an effective solution for improving clustering efficiency.
• The integration of diverse base learners and advanced ensemble strategies leads to superior performance.
• The method shows promise for diverse data clustering applications.