Description: Training-test data set with 75%-25% ratio and model performance measurement using f1 score, precision and recall metrics. Improve results by experimenting with input parameters. Training SVM with the best parameters and listing findings on how much the classification quality was affected. Use of neural networks.