A Research Survey on State of the art Heart Disease Prediction Systems

Abstract

Disease prediction systems are the better alternatives, to avoid the human errors in disease diagnosis and also assist in disease prevention with early detections. Highdemand in preventing the rapidly increasing heart disease death tolls expanded the horizons of the former research scholars for introducing the intelligent heart disease prediction systems. Prediction of heart disease from patient's health record attributes is, a proven multi-dimensional decision-making system, which merely depends on mining attribute correlations too. Patient Health Records (PHR's) with structured categorical data and unstructured text/image data are the major input resources for heart disease prediction. Heart disease dataset preparation, prediction system's process flow design, process execution and results evaluation are the most common life cycle modules of any heart disease prediction system. Although many former research were introduced various heart disease prediction models, but they are still suffering from some common set of problems. Input dataset attributes modeling, attribute risk factor calculation, correlations mining; threshold determination and achieving the high accuracy in disease prediction are the major limitations of the existing heart disease prediction systems. As part of my research on designing intelligent heart disease prediction models, several research papers are analyzed and narrated that knowledge in a proper manner with detailed description. The main objective of this study is to represent the current scenario of heart disease prediction systems and their associated modules in brief.