Accurate and Fast Diagnosis of Heart Disease using Hybrid Differential Neural Network Algorithm

Abstract

introduced the system that deals with Thyroid diseases [11]. The data has been collected from the UCI repository and undergo pre-processing. The data which had pre-processed is multivariate in nature. Using Hybrid Differential Evolution based algorithm the Dimensionality has been followed to optimize 21 to 10 number of attributes. The subset has been provided to Support Vector Machine classifier algorithm. Then in order to stabilize the errors this process has been carried iteratively and data is classified. Thus accuracy of classification has been observed. The authors of [12] proposed a system that analyzes the fuzzy neural networks uses and implementations based on FPGA in order to predict various physiological diseases. This Strategy works with path physiological data that has been collected. Then Fuzzy Rule are prepared as per the guidelines of doctors. Finally using Neural Network map these rule in order to predict the specific disease. This process has been carried out based on predefined parameters. FPGA had implemented using Fuzzy-Neural algorithm which is used as an expert system in the field of disease prediction. Isaac Trigueroet al., proposed a methodology for optimize the position of prototypes relay on differential evolution approach [13]. They made various study regarding performance of recent things in differential evolution. This process has been carried out by optimizing the position which is prior to nearest neighbor classification. Thus results have been identified. The authors of [14] brought out the system for diagnosis the heart disease which is based on the ReliefF and Rough Set method by sing hybrid classification system. This system contains RFRS feature selection and a classification along with an ensemble classifier. These include data discretization, feature extraction and feature reduction. To deal with these authors had developed the algorithm called heuristic Rough Set reduction. On the other hand, an ensemble classifier has been proposed depends on C4.5 classifier. The dataset on heartis collected from UCI database. Finally system achieved the maximum classification accuracy of about 92.59%. The authors of [15] used the Bat Imperialist Competitive Algorithm which is based on an evolutionary algorithm on the development of human's socio-political. In this initial population has been considered with m features and n cardiovascular heart disease observations. This approach helps to minimize number of features that indirectly reduce the diagnosis tests mandatory for patients. This system acquires 94.4% accuracy and provides better accuracy for diagnosis cardiovascular heart disease. In [16] authors proposes a method use Differential Evolution (DE) for hybrid classification algorithm and Least Squares Support Vector Machine which is used for classification. DE algorithm is using the parameters that guarantee the effectiveness. This algorithm evaluates Breast Cancer dataset collected from UCI Repository. This algorithm pays a way for comparing with different classifier based algorithms on same database. Thus it achieves an accuracy of 99.75%. Alsalamahet al., aims to develop an informatics system for classification of heart diseases with the help of data mining techniques for Radial Basis functions and emerging Neural Networks approach [17]. The author deals with various processes such as classification system using Radial Basis neural networks for Coronary Artery Disease. Then by deep learning approach different types of heart diseases has been identified. Thus it provides novel technique by using Neural Networks techniques, and it evaluates the accuracy and performance by compare the results of different classification models. In [18] the authors proposed a technique for artificial neural system learning by differential transformative calculation. DE with global and local neighborhood-based change (DEGL) calculation scans for synaptic weight coefficients and to limit the learning mistake. Both global and local neighborhood-based mutation operator have joins to shape contributor vector. This strategy works for order of certifiable information and along these lines results drew out the proficiency and viability. The authors of [19] gave writing on different optimization algorithms, machine learning algorithms and applications in smart healthcare. This paper additionally talked about with difficulties, security, pilot studies and genuine venture and correspondence between medicinal staffs and information investigation which are basic for savvy human services. This proposition clears an approach to guarantee the wellbeing administrations accessible in future that are earth supportable and monetarily reasonable. In [20] the author utilizes the last populace of heterogeneous flexible neural trees (HFNTs) or, in other words Pareto-based multi-objective genetic programming (MOGP). The parameter tuning by differential development draws out the astounding gathering frameworks. Enhancement of exactness, decent variety and intricacy gives answer for auxiliary multifaceted nature. In this manner chose hopefuls helps as great outfit framework. These methodology backings to better execution over the calculations that are gathered from correlation. Consequently this paper demonstrates that HFNT is viably utilized for data analysis and modelling.