Data mining in soft computing framework: a survey

Area of Science:

• Computer Science
• Artificial Intelligence

Background:

• Data mining involves extracting valuable information from large datasets.
• Soft computing methodologies offer robust approaches to handle complex and uncertain data.

Purpose of the Study:

• To survey the literature on data mining using soft computing.
• To categorize soft computing tools and their applications in data mining functions.

Main Methods:

• Literature review and categorization of soft computing tools.
• Analysis of fuzzy sets, neural networks, genetic algorithms, and rough sets.
• Evaluation of hybridization techniques in soft computing for data mining.

Main Results:

• Fuzzy sets excel in understandability, handling incomplete data, and human interaction.
• Neural networks offer robust learning and generalization in data-rich environments.
• Genetic algorithms provide efficient model selection, and rough sets manage data uncertainty.

Conclusions:

• Soft computing methodologies are highly effective for diverse data mining tasks.
• Different soft computing tools offer unique advantages for specific data mining challenges.
• Further research is needed to address challenges in applying soft computing to data mining.