Behavioral Prediction of Cancer Using Machine Learning
Ashish Kumar, Rishit Jain
Abstract
In the past few decades, the medical and technical industry has made major strides in improving the diagnosis and prognosis of high- and low-risk cancer, as well as the different types of cancer. As precise as the diagnoses may be, the prognosis of a patient is often seen as a fortuity that the condition can be dealt with effectively or not. Millions of people around the globe are diagnosed with some type of cancer every year. Pathologists have been performing cancer diagnoses and prognoses for decades, but it still causes millions of people to face uncertainty about their fate. Researchers have made major strides in the field of pathology to identify and attack the disease before it runs amok. But the prognoses and treatment procedures are still not nearly as accurate as the diagnoses can be. It is about time that crucial steps are taken to develop the norm of pathology and help save the millions of people who live every day under the threat of death. Machine learning (ML) could be the next leap in the development of pathology. ML can facilitate the process of an early diagnosis and prognosis of the malignancy of different cancers, as it can ultimately ease the subsequent clinical treatment of affected patients. Several teams of professionals from biomedical and bio-informatic fields have conducted research and used different ML tools to detect the primary features from complex datasets to classify the disease. Several ML techniques such as support vector machines (SVMs), artificial neural networks (ANNs), Bayesian Networks (BNs), and even decision trees (DTs) have been employed to tackle the problem of an automated diagnosis of cancer. Even though ML clearly poses as a solution for the early diagnosis problem, it still requires validation up to a certain extent for these techniques to be used for clinical practice. Through this research, we aim to present a comparative review of the different ML tools that can be used for cancer prediction and to find out which approach presents the most potent results.