ANN based Expert System to Predict Disease in Cardiac Patients at Initial Stages

摘要

Objective of this research is to develop an expert system for the preliminary investigation of cardiac abnormality in human beings. Artificial Neural Network (ANN) is judged best for the prediction of heart abnormalities in cardiac patients at initial stages. Our research is intended to employ an Artificial Intelligence (Al) technique in an automated solution, having minimum error bounds. An ANN based expert system is designed and developed, which identifies presence or absence of cardiac disease in patients by considering best practiced disease symptoms. The proposed expert system may help the clinicians in the preliminary investigation of cardiac abnormality in human beings. Artificial Intelligence (AI) techniques have been extensively utilized in the design and development of expert systems (Mitchell, 1997; Cowan, 2005; Jackson, 1990). They have applications in the mining of large datasets (Phyu, 2009). Patterns and trends can be learned and may be utilized in the decision making process (Dashti et al., 2010). Expert systems have applications in domains like remote sensing, simulations, and disease predictions (Pal, 2007; Ivezic & Gar-rett, 1998; Veropoulos, 2001; Oh et al., 2009). Most common and widely used AI techniques include but not limited to Artificial Neural Network's (ANN's), Decision Tree's (DT's) and Naive Bayes' Classifier's (NBC's) (Mitchell, 1997). AI techniques have been widely explored for medical domain in the disease prediction. They are employed in the prediction of orthopedic problems (Mantzaris et al., 2008), lung cancer (Mughal & Ikram, 2004), diabetes (Huang et al, 2004; Sopharak et ah, 2010), and cardiac abnormalities (Tantimong-colwat et al., 2008; Qazi et al, 2007; Yaghouby et al., 2009; Parthiban & Subramanian, 2008; Patii & Kumaraswamy, 2009; Adams & Choi, 2012; Srinivas et al., 2010).