This project presents investigation and results of multi-class classification using Linear Discriminant Analysis (LDA) and Support Vector Machines (SVM). Furthermore,  by using different derivatives of LDA and SVM and different classification approaches, we obtain confusion matrices that describe the performances of each technique. Finally, a table summarizing the classification accuracy of each technique shows that SVM with Gaussian kernel, after PCA projection, using OVA classification method, achieves the highest classification accuracy.