| Popis: |
Mustafa Jan,1 Akber A Awan,2 Muhammad S Khalid,1 Salman Nisar1 1Department of Industrial and Manufacturing Engineering, Pakistan Navy Engineering College, National University of Sciences and Technology, Karachi, Pakistan; 2Department of Management and Information System, Pakistan Navy Engineering College, National University of Sciences and Technology, Karachi, Pakistan Abstract: In health care informatics, the predictive modeling solution for cardiovascular risk estimation is extremely challenging. Thus, the attempt to clinically screen the medical databases and predictive modeling through soft computing tools is regarded as a valuable and economical option for medical practitioners. Therefore, the soft computing tools are today’s need in health care application, which can perform data analysis and modeling, and they can assist the physician to make right and prompt clinical decisions. Extracting patterns that tie predictor’s variables in a health science database is the topic of data mining. Existing data mining techniques are appropriate to model complex, dynamic processes. In this work, we propose Ensemble model approach for combining the predictive ability of a multiple classifiers’ model for better prediction accuracy. In this study, ensemble learning combines five classifiers’ model approaches, including support vector machine, artificial neural network, Naïve Bayesian, regression analysis, and random forest, to predict and diagnose the recurrence of cardiovascular disease. The cardiovascular data records of Cleveland and Hungarian were extracted from the UCI repository. Experimental results demonstrated that the ensemble model is a superior approach in terms of high predictive accuracy and reliability of diagnostics performance. In addition to this, this study also presents a smart heart disease prediction system as a valuable, economical and prompt predictive option having friendly graphical user interface, which is scalable and expandable. Keywords: ensemble methods, smart heart disease prediction system, data mining model, classification techniques |
