Finding Structure in Time Series
Clustering, Ranking, Recommendation and Anomaly Detection
Volume 1: Clustering, Ranking, Recommendation and Anomaly Detection
Finding Structure in Time Series
Machine learning methods for time series, worked on energy and supply chain data, built to be run, not just read.
Time Series
Data & ML
Decision
The arc every chapter follows
Structure of the book
The definition and validation of load-profile clusters, a pipeline validated on three utilities, the reliability of customer archetypes, and what clustering contributes to network risk under rooftop PV and EV charging.
Retrieval of the risk of a new connection, ranking of transformers for reinforcement and of live violations for operators, and the recommendation of mitigation measures, evaluated against simulated network outcomes.
Detection of faulty meters, non-technical loss and inverter faults, and of violations that depart from their violation pattern, with thresholds whose false-alarm rate is guaranteed.
- 01Foundations of Anomaly Detection
- 02Meter Data Quality and Customer Behaviour
- 03Non-Technical Loss and Changes over Time (in preparation)
- 04Anomalous Violations on the Network (in preparation)