Heart Disease Classification Using Multiple K-PCA and Hybrid Deep Learning Approach

One of the severe health problems and most common types heart disease (HD) is Coronary (CHD). Due to lack a healthy lifestyle, HD would cause frequent mortality worldwide. If attack occurs without any symptoms, it cannot be cured by an intelligent detection system. An effective diagnosis CHD should prevent human casualties. Moreover, systems employ clinical-based decision support approaches assist physicians in providing another option for diagnosing detecting HD. This paper aims introduce prediction model including phases like (i) Feature extraction, (ii) selection, (iii) Classification. At first, feature extraction process carried out, where features time-domain index, frequency-domain geometrical domain features, nonlinear WT signal energy, skewness, entropy, kurtosis are extracted from input ECG signal. The curse dimensionality becomes issue. provides solution this issue introducing new Modified Principal Component Analysis known as Multiple Kernel-based PCA reduction. Furthermore, dimensionally reduced set then subjected classification process, hybrid classifier combining both Recurrent Neural Network (RNN) Restricted Boltzmann Machine (RBM) used. last, performance analysis adopted scheme compared over other existing schemes terms specific measures.