Egyptian Computer Science Journal (ISSN-1110-2586) Volume 41– Issue 2, May 2017 Analysis of Electrocardiogram for Heart Performance Diagnosis Based on Wavelet Transform and Prediction of Future Complications 1 Kamel H. Rahouma, 2 Rabab H. Muhammad, 1 Hesham F.A. Hamed, 1 Mona A. Abo Eldahab 1 Electrical Engineering Dept., Faculty of Engineering, Minia University, Minia, Egypt 2 Computer and System Engineering Dept., Faculty of Engineering, Minia University, Minia, Egypt kamel_rahouma@yahoo.com Abstract This paper aims to analyze the heart electrocardiograph (ECG) to diagnose the heart performance and predict any future complications. The authors utilize the MIT BIH database to obtain the original data. The discrete wavelet transform (DWT) is used to decompose the ECG signal and to reconstruct it. The means of different performance measures of the heart performance (QRS,RR,ST,QT,PR) are tested for their significance before comparing them to the limits of normal performance. Any abnormal measure is used to diagnose the corresponding disease(s). Prediction techniques are used to expect any future complications regarding the abnormal measure(s). Three methods of prediction are applied: the Linear Prediction Method (LPM) and the Grid Partitioning and Fuzzy c-mean Clustering based on Neuro-Fuzzy prediction. Keywords: Electrocardiograph (ECG), Heart Diseases Diagnosis, Wavelet Transform, Linear Prediction Method, Grid Partitioning Method, Fuzzy c-mean (FCM),Neuro-Fuzzy (ANFIS). 1. Introduction For the time being, the computerized systems are one of the most important systems for analyzing the biomedical electrical signals. This has become very important especially for understanding the real biological processes such as the electrical activity signal of the heart (i.e., Electrocardiography or ECG) or the electrical activity signal of the brain (i.e., Electroencephalograph or EEG) or the electrical activity signal of the muscles (i.e., Electromyography or EMG). Computerized techniques have many advantages in features extraction of the biomedical signals and this helps in diagnosing the diseases of the mentioned parts [1]. In this paper, we are interested in diagnosing the heart diseases. Diagnosing the heart diseases depends on processing its beats information to extract some main features which are used in the diagnosis. These features represent the bioelectric signal generated by the processes of the repolarization and depolarization of the heart. Thus, the accurate detection of ECG depends on the values of a trial depolarization (P), the distance between the start of a trial depolarization and the start of ventricular depolarization (PR), the ventricular depolarization (QRS complex), the temporary pause of the ventricular electrical activity before repolarization (ST segment), the ventricular repolarization (T wave) and the -11-