Early Prediction and Diagnosis for Cancer Based on Clinical and Non-Clinical Parameters: A Review

Abstract

Cancer seems to be a heterogeneous disease that comprises of numerous different subtypes. Predicting and finding cancer at an early stage will be helpful to fully eradicate the disease. Data mining is a potential solution to generate the rules from vast datasets that are used in medical domain. In this paper, a comparative analysis of existing prediction models in medical domain has been done. Some of the most quoted paper in the field of medical science is listed along with their methodologies, merits and demerits in a table. Classification techniques and the evaluation metrics used to compare the prediction models are also explained. Some of the tools used for machine learning are also discussed. This paper can give a good heads up to the one who wants to use predictive analysis for medical applications.