A Neighborhood Rough Sets-Based Attribute Reduction Method Using Lebesgue and Entropy Measures.

Area of Science:

• Data Science
• Machine Learning
• Rough Set Theory

Background:

• Traditional attribute reduction methods struggle with continuous data, leading to low accuracy and high cardinality.
• Effective attribute reduction is crucial for enhancing classification performance in complex datasets.

Purpose of the Study:

• To propose a novel attribute reduction method for continuous numerical data using neighborhood rough sets.
• To address limitations of traditional methods in handling continuous data and maintaining classification information.
• To improve classification accuracy and reduce computational complexity for high-dimensional data.

Main Methods:

• Fisher score method for irrelevant attribute elimination.
• Introduction of Lebesgue measure into neighborhood rough sets for uncertainty quantification.
• Development of neighborhood entropy-based uncertainty measures and neighborhood roughness joint entropy.
• Design of a heuristic attribute reduction algorithm.

Main Results:

• The proposed method effectively handles continuous numerical data.
• Demonstrated ability to maintain original classification information.
• Significant reduction in computational complexity for high-dimensional data.
• Experimental results show high classification accuracy on various datasets.

Conclusions:

• The novel attribute reduction method using Lebesgue and entropy measures is effective for continuous data.
• The approach enhances classification performance and reduces complexity in large-scale datasets.
• The study provides insights into the essence of knowledge and uncertainty in neighborhood decision systems.