Medical data set classification using a new feature selection algorithm combined with twin-bounded support vector...

Area of Science:

• Medical Informatics
• Machine Learning
• Data Mining

Background:

• Early disease diagnosis and treatment are critical for preventing mortality.
• Data mining and machine learning offer valuable tools for improving diagnostic accuracy and reducing errors.

Purpose of the Study:

• To introduce a novel feature selection algorithm.
• To evaluate the efficacy of the proposed algorithm in medical data classification.

Main Methods:

• Developed a new feature selection algorithm.
• Combined the feature selection method with a twin-bounded support vector machine (FSTBSVM).
• Validated the approach using eight benchmark medical datasets and 10-fold cross-validation for parameter optimization.

Main Results:

• The proposed feature selection method combined with FSTBSVM demonstrated high efficiency in medical data classification.
• The FSTBSVM achieved robust performance, validated by classification accuracy, sensitivity, and specificity analyses.
• The method successfully reduced the number of features required while maintaining good results.

Conclusions:

• The novel feature selection algorithm integrated with FSTBSVM is a promising technique for medical data classification.
• This machine learning approach can enhance diagnostic capabilities by providing accurate results with reduced feature sets.
• The findings support the potential of FSTBSVM for early disease detection and improved patient outcomes.