| Popis: |
Most of the developed countries are facing the threat of rapid growth in cases of cancer patients. Cancer is one of the diseases that lead to a high number of deaths every year. Cancer treatment efficacy in medical practice majorly depends upon prediction of cancer susceptibility, cancer prognosis, and cancer recurrence. In recent years, statistical approach of machine learning techniques is proven to be a boon for diagnosis, classification, prediction, and prognosis purposes in health care. Various researchers are implementing machine learning concepts to improve cancer prediction and prognosis. Decision trees, KNN, SVM, and Neural networks are applied to predict the survivability of cancer patients with high accuracy. Recently, researchers are more focused on Deep Learning for better results. However, features for cancer prognosis vary with the type, site, and stage of cancer. For oncological statistical studies, histopathological, biomarker, and genes expression profile data are analyzed, and relationships are inferred for cancer prediction and prognosis. A majority of the studies have been carried out in breast, prostate, and lung cancer as these are most prevailing cancer types. With the emerging technological hazards, environmental factors, and ill habits of people, other types of cancer are also becoming more and more common. The present study provides a review of various features for cancer prognosis and accuracy in prediction of cancer prognosis using different machine learning techniques. |
