Research Article
Heart Disease Diagnosis Using Reconstructive Radial Basis Function Networks with Overlapping Prevention Method
@INPROCEEDINGS{10.4108/eai.14-10-2015.2261993, author={Mashail Al-Salamah and Saad Amin}, title={Heart Disease Diagnosis Using Reconstructive Radial Basis Function Networks with Overlapping Prevention Method}, proceedings={5th EAI International Conference on Wireless Mobile Communication and Healthcare - "Transforming healthcare through innovations in mobile and wireless technologies"}, publisher={ACM}, proceedings_a={MOBIHEALTH}, year={2015}, month={12}, keywords={artificial intelligence data classification heart diseases radial basis function networks}, doi={10.4108/eai.14-10-2015.2261993} }- Mashail Al-Salamah
Saad Amin
Year: 2015
Heart Disease Diagnosis Using Reconstructive Radial Basis Function Networks with Overlapping Prevention Method
MOBIHEALTH
ICST
DOI: 10.4108/eai.14-10-2015.2261993
Abstract
The term “Heart disease” applies to any abnormal heart condition affecting the heart itself or the blood vessels. It is prevalent today as it is the leading cause of deaths in the U.S. Because the heart is the engine of blood circulation, any issue it suffers directly affects the whole body. In this paper, the design and implementation of a Radial Basis Function Network is presented to interpret, via data classification, the diagnoses of heart disease patients. In the designed classifier, the Gaussian distribution function is used as a kernel of RBFs to build up the network, and peak RBF values are determined between -100% and +100% according to whether a patient has a certain disease or not. This creates smooth gradients between different RBFs, allowing the network to act as a fuzzy system. The designed classifier has special training and optimisation algorithms, with those it aims to use the classification space at its maximum potential. This classifier is implemented as a standalone computer software. An experiment using the designed system on two datasets collected from Prince Sultan Cardiac Center, Saudi Arabia, and UCI Machine Learning Repository, achieved good results.
