Machine-Learning

• Developed a software system to identify six universal basic facial expression based on labeled dataset via Matlab programming
• Trained supervised machine learning models including Artificial Neural Networks (ANN), Decision Tree (DT), and applied ID3 algorithm and pruning skill to optimize the classifier, as well as regularization with optimal parameters to overcome overfitting
• Evaluated model performance of classification via k-fold cross-validation technique, confusion matrix, recall and precision rates and measure
• Visualized model training results by dimensionality reduction using Correlation Based Feature Selection (CFS) and Principal Component Analysis (PCA), in this way to simplify the models through the removal of bad features