Affiliations: [a] Tata Consultancy Services, Bengaluru, Karnataka, India
| [b] School of Computing, SASTRA Deemed University, Thanjavur, Tamil Nadu, India
| [c] Sri Ramachandra Faculty of Engineering and Technology, Sri Ramachandra Institute of Higher Education and Research, Chennai, Tamil Nadu, India
Correspondence:
[*]
Corresponding author. G. Manikandan, School of Computing, SASTRA Deemed University, Thanjavur, Tamil Nadu, India. Email: manikandan@it.sastra.edu.
Abstract: Heart diseases are one of the crucial diseases that may cause fatality in both men and women. About 12 million deaths occur across the world due to heart diseases. With the advancement in information technology, it is possible for the Healthcare industry to store enormous volume of data containing millions of patient’s medical information along with their treatment details. If utilized in an efficient manner, this information helps the doctors to diagnose the diseases in a precise manner. Data mining algorithms are employed to analyse huge data sets and to discover unseen patterns. Data mining plays an essential role in medical diagnosis. Doctors bank on different computer models which uses data mining algorithms to prefigure different kinds of diseases in patients. So, the need is to design a methodical data mining algorithm that helps for better forecast of diseases. The main goal of this work is to create an ensemble of algorithms which results in better accuracy. The ensemble is constructed by making use of stacking ensemble technique, which comprises of two categorization algorithms namely Naïve Bayes and Artificial Neural Network. The Cleveland heart disease data set acquired from UCI machine learning repository containing 14 attributes and 303 instances is given as input to these algorithms. From our experimental analysis it is evident that the proposed ensemble scheme results in a better accuracy.
