Classification and Detection of Mesothelioma Cancer Using Feature Selection-Enabled Machine Learning Technique

Area of Science:

• Oncology
• Medical Informatics
• Data Science

Background:

• Malignant mesothelioma (MM) is a rare, aggressive cancer often diagnosed at advanced stages, with limited treatment efficacy.
• A strong association exists between asbestos exposure and the development of malignant mesothelioma.
• Current treatments for advanced MM show limited success, highlighting the need for improved diagnostic and classification methods.

Purpose of the Study:

• To propose a novel classification and detection method for malignant mesothelioma (MM) utilizing machine learning.
• To enhance the accuracy of MM classification through effective feature selection techniques.

Main Methods:

• Employed the CFS (correlation-based feature selection) approach for identifying relevant features for MM classification.
• Utilized Naive Bayes, Fuzzy SVM, and the ID3 algorithm for classifying mesothelioma cancer.
• Evaluated the performance of machine learning strategies using various metrics.

Main Results:

• Feature selection using CFS demonstrated a substantial positive impact on the accuracy of the classification models.
• The chosen machine learning algorithms (Naive Bayes, Fuzzy SVM, ID3) were effective in classifying mesothelioma.
• The selection of pertinent features was found to be critical for improving classification accuracy.

Conclusions:

• Machine learning, particularly with effective feature selection, offers a promising avenue for improving malignant mesothelioma classification and detection.
• The CFS approach is effective in identifying key features that enhance the performance of mesothelioma diagnostic models.
• Further research into machine learning applications can aid in earlier and more accurate diagnosis of this rare cancer.