Optimal Feature Selection and Classification with Firework Optimization for Heart Attack Prediction by Machine Learning Classifiers

Evolutionary optimization has rapidly emerged as a prominent field of research and application in computational intelligence. Machine learning and Deep learning are subset of artificial intelligence that provides an essential stage in the preprocessing phase involving feature selection. Numerous approaches have been developed for feature extraction and selection, but one of the most effective methods for identifying optimal features is through the application of an optimization algorithm. In this work, the heart attack prediction dataset is considered to identify the optimal features. Since many factors influence the cause of a heart attack, it becomes crucial to choose the optimal features that have a significant impact. The manual approach of selecting the features is time-consuming when the number of dimensions is high. Thus, in this work, an evolution-based optimization approach using Firework Optimization is proposed with five different machine learning classifiers. The performance is measured by applying the various classifiers. Gaussian Naïve Bayes proved better compared with Random Forest, Support Vector Machine (SVM), Decision Tree, and K-Nearest Neighbor, as its computation time is less compared with other classifiers. Based on the optimal features selected, further classification is done, and performance is measured for each classifier.