Comparative Analysis of Multiclass Heart Disease Prediction Classification Models using Preprocessing and Feature Selection

Prediction of heart disease and diagnosing it is a challenging task in the medical industry. All over the world, heart disease is considered as the deadliest disease in the life of the human being. Most deaths occur if the cause of the disease is not known in advance. The correct prediction of heart disease can save the life of the person. By selecting significant features and applying classification algorithms, heart disease could be predicted in an efficient way. In this paper, the preprocessing techniques like handling missing values, eliminating duplicate rows, and removing outliers have been applied to the UCI repository heart disease dataset. Machine learning classification algorithms such as SVM, Gradient Booster, Random forest, and Decision Tree were implemented in the dataset and the prediction accuracy was compared. The result shows that these models have produced more accuracy after preprocessing and feature selection techniques.