Abstract
A recent study shows that almost 30% of total global deaths are caused by
heart disease. These days precise diagnosis related to heart disease is very difficult. The
doctor advises patients to take various tests for diagnosis, which is a very costly and
time-consuming process as medical databases are large and cannot be processed
quickly. A new approach has been proposed to predict heart disease from historical
data sets. In this chapter, heart disease possibilities in patients are predicted with the
help of neural networks on distributed computing. Feature selection was applied to the
dataset to get better results and to increase the performance. Feature selection reduces
the number of attributes from the dataset and only provides the necessary attributes,
which directly reduces the number of tests required for the diagnosis.