Model for Predicting Cardiac Health using Deep Learning Classifier
Richa Sharma, Shelly Gupta, Puneet Garg
Abstract
Data mining in medical domain is the most active research area and mainly concerned of discovering hidden pattern and features within data, deep learning involves mainly a neural network with large number of hidden layers for difficult machine learning task. Data mining includes deep learning and would help to make sense from data. In this research paper a model is generated by using deep learning classifier for correct classification of heart disease. For this study data was collected by retrospective method. It includes 9 attributes (including predictive attribute) and 209 instances. This paper firstly identified that the people having diabetes and hypertension are the most dominant category of heart patients, secondly classification was done by comparing old data mining algorithm with deep learning and this paper also shows the comparative analysis of different types of neural network with deep learning algorithm. The statistical testing result indicated that the maximum accuracy of 71.4% was achieved using deep learning, parameters like precision, recall/sensitivity, specificity, F measure were calculated for further investigation. Model was generated using RapidMiner version 8.2.