| Summary: | Malignant mesothelioma is a rare but aggressive kind of cancer that mainly occurs in the mesothelial surfaces of the lung or chest cavity, abdomen, or other internal organs. Early diagnosis of the disease leads to applying a safer and more effective treatment for the patient. Nowadays, applying machine learning techniques to diagnose and diagnose different diseases has become a common approach providing more efficient, cost-effective, and fast screening of the disease. The current study proposes a model namely BHHO+DT to diagnose mesothelioma through computational techniques using patient health records. The proposed method first employs a binary Harris hawks optimization algorithm to select highly influential features in disease diagnosis, and then, uses the CART decision tree classifier on the selected features to train the model. The importance of the selected features was confirmed by a Gini impurity decrease analysis study. The experimental results reveal that the method obtains an accuracy higher than 85% for the balanced dataset. This accuracy was obtained after removing invalid features from the dataset. As a result, the high applicability and preference of the proposed method are demonstrated in terms of accuracy in comparison to other machine learning methods.
|
|---|