An Efficient Classification Approach for Predicting Cause of Death using Mixed Probability Rule Based Algorithm

摘要

Now a day’s examining the health of each person in the every country is an integral part of healthcare. After examining the health of each person we can identify type of risk to be occurred. The analysis of risk based unlabelled data can be done by using classification approach in the data mining. Particularly we are take unlabelled data contains information related to participants in the health examination whose health condition is vary from great health to very ill. In this study we formulated the task of risk prediction as a multiclass classification problem using the Cause of Death (COD) information as labels, regarding the healthrelated death as the “highest risk”. The goal of risk prediction is to effectively classify 1) whether a health examination participant is at risk, and if yes, 2) predict what the key associated disease category is. In other words, a good risk prediction model should be able to exclude lowrisk situations and clearly identify the highrisk situations that are related to some specific diseases. In the examination of health we are identifying different states of health without ground truth. So that by predicting risk of each participant by using classification approaches in the data mining. In this paper we proposed Mixed Probability Binary Rule Based Classification Algorithm is used to predict health risk of participate person. By implementing this algorithm we can get efficient classification result and also give better performance.