CLASSIFICATION OF HEART DISEASE RISK WITH MULTILAYER PERCEPTRON: A MACHINE LEARNING APPROACH

Abstract

This research explores the integration of Principal Component Analysis (PCA) with Support Vector Machine (SVM) classification to identify the key factors influencing customer satisfaction using questionnaire data. Leveraging SVM's effectiveness in questionnaire-based classification and PCA's feature reduction capabilities, the study examines the impact of PCA on model performance. From a dataset comprising responses from 100 participants regarding product quality, price, and service level, PCA was employed to reduce the questionnaire features to three components. Results revealed Product Quality as the most significant factor affecting Customer Satisfaction, while Service Level exhibited the lowest influence. Comparative analysis of SVM models with and without PCA integration demonstrated improved performance metrics, including reduced error rates and enhanced accuracy. PCA mitigated overfitting risks, enhancing model generalization and interpretability. The findings underscore the practical significance of PCA in optimizing SVM classification models for customer satisfaction analysis. Future research may focus on optimal component selection and the robustness of PCA-SVM models across diverse domains, advancing machine learning applications in real-world scenarios.

Keyword : Classification, Customer satisfaction analysis, Feature reduction, Principal Component Analysis, Support Vector Machine