Improved Heart Disease Prediction Using Deep Neural Network

Heart disease is biggest challenge for medical professionals. Modern life style made it an epidemic; according to a survey conducted by WHO heart attack is leading cause of death all over the world especially in the western world. It is surveyed that 23% of the death in US is due to Heart related disease [1]. It has been observed assistance is needed for helping medical professionals in detecting the chance of heart attack in the human. In recent times a lot of work related to providing an automated support system for predicting chance of Heart attack in human has been done. After advancement of computer science, researchers felt that they can help in some of the key interdisciplinary areas like medical science. Machine learning techniques are compared on the single data set which does not reflect true potential of any algorithms. They also suffer from some of the key anomalies such as accuracy and manual data set pre-processing. In this paper, we propose Deep Neural Network methods for creating an automated system for heart attack prediction. It is tested on multiple dataset to find out true potential and providing certainty in the accuracy. Method also promises to remove all the mentioned anomalies from the system like lack of accuracy and automated approach in pre- processing of the data set. In result analysis, it has been observed that prediction is much more efficient and minimum accuracy achieved through this proposed method is 87.64% on any of the data set taken under consideration.