• Developed an algorithm to detect anomalies in a time-series dataset sourced from a city infrastructure system, where anomalies indicated cyberattacks.
• Utilized a training dataset containing only nominal instances and analyzed a testing dataset with both nominal and anomalous cases.
• Identified deviations from learned nominal patterns to detect anomalies, using test labels to indicate nominal (0) or anomalous (1) instances.
• Employed advanced Machine Learning models and techniques such as KNN Imputer, DecisionTreeClassifier, OneClassSVM, and Isolation Forest to ensure precision and accuracy in anomaly detection.
• Applied dimensionality reduction techniques like PCA and FastICA to enhance model performance.