Graphical User Interface-Based Multi-Class SVM Classification Experimental System

This experimental system implements a multi-class classification approach using Support Vector Machine (SVM) algorithms through an intuitive graphical user interface. The entire system is developed using MATLAB programming language, incorporating core functions for data handling, feature extraction, and SVM model optimization using libraries like fitcsvm for binary classification extended to multi-class scenarios through one-vs-all or one-vs-one strategies. The system represents a supplementary program for my graduation research project, playing a significant role in my academic investigations. Beyond the current SVM implementation, the architecture supports seamless integration of additional classification algorithms through modular function interfaces, allowing extensions such as k-NN, decision trees, or neural networks by implementing corresponding training and prediction modules. We envision this experimental system serving as a convenient and user-friendly tool for researchers and students conducting multi-class classification experiments. The interface provides interactive controls for parameter adjustment, real-time results visualization through confusion matrices and ROC curves, and export capabilities for analysis reports - making it particularly suitable for educational demonstrations and comparative algorithm studies.