Comparative analysis of machine learning-based approaches for prediction of colon cancer.

Colon cancer, along with lung, prostate, and breast cancer, is one of the most common types of cancer globally and is regarded as one of the world's leading causes of death. In therapy for this common illness, substantial progress has been achieved recently. Adjuvant chemotherapy is effective, specifically in stage III patients, and surgery was tailored for the most outstanding results with the least amount of morbidity. Several novel target-oriented medications are being evaluated, and some of them (cetuximab and bevacizumab, for example) have already shown high activity/efficacy, primarily when used in conjunction with chemotherapy. To get the best results, it is necessary to provide current recommendations for treating these patients in medical trials and daily practices. This research seeks to use a person's gene expression to determine whether or not they have colon cancer. This information is used by the Logistic Regression, Naive–Bayes Algorithm and Catboost Algorithm to predict human colon cancer. The learning classification dataset included 7457 gene expression levels from 36 patients, 18 of whom were healthy and 18 of whom were malignant. [ABSTRACT FROM AUTHOR]