Repository for "Data Mining - Advanced Topics and Applications" project exam.
This project consists in analysing and processing audio signals, employing advanced data mining/machine learning algorithms on the FMA dataset.
This analysis is focused on:
- Imbalanced learning: Random Undersampling, CNN, Tomek's Link, Random Oversampling, SMOTE, K-Means SMOTE, ADASYN
- Anomaly detection: DBSCAN, KNN, LOF, ABOD, Isolation Forest, Extended Isolation Forest, Autoencoders;
- Advanced classification methods: Naive Bayes, Rule-based classifiers, Logistic Regression, SVM, Ensembles (Random-Forest, Bagging, Adaboost) and Neural Netoworks (MLP);
- Time series analysis: Motifs & Anomaly detection, Clustering, Shaplet-based classifiers;
- Sequential Pattern Mining and Advanced Clustering: X-Means, OPTICS and Transactional Clustering (K-Modes);
- AI Explaianbility: LIME explainer;
Dataset and additional info are available at: mdeff/fma