Diagnosing heart disease by a novel hybrid method: Effective learning approach

Abstract

In this paper, the effective classifier Extreme Learning Machines (ELM), Enhanced Fast Learning Networks (EFLN), Support Vector Machines (SVM), and Decision Trees (DT) for early Heart Disease (HD) diagnostics have been investigated. Using the optimal parameters for the proposed method, ELM, EFLN, SVM, and DT are considered for early diagnosis of heart disease. The performance of the proposed method has been evaluated versus the Accuracy (ACC), Sensitivity (SE), and Specificity (SP) HD dataset. This study also proposed a new effective Hybrid Model, by combining the Particle Swarm Optimization Algorithm (PSOA) with the most effective Learning classifiers. The proposed hybrid diagnostic method records ACC through ten runs for a 10-fold cross-validation (CV). According to the results, the proposed method shows an ACC of 93% for (PSO-ELM), 96% (PSO_EFLN), 91% (PSO-SVM) and 93% (PSO-DT).

Keywords

• Heart disease (HD)
• Classification
• Extreme learning machine (ELM)
• Particle swarm Optimization algorithm (PSO)
• Hybrid models
• Enhanced fast learning network (EFLN)