Anomaly detection
Anomaly detection helps in the early detection of critical outliers in a system. Based on the context, these outliers can be detrimental and result in loss of resources, and time through errors, fraud, manipulation of stocks, and other such malicious activities. Outliers can also be beneficial in investing, and arbitrage. Business decisions that leverage anomaly detection, which used to require intense human resource and capacity can now be completed in a short time through versatile models and automation.
Projects goal
L1: Feature Engineering for various datasets that can capture data anomalies.
L2: Critique potential usefulness of features towards anomaly detection.
L3: Develop models for advanced supervised and unsupervised learning.
L4: Execute automated machine learning using commercial or open source packages.
L5: Recommend strategies for monitoring model performance once in production.