This Repo contains some Poc level code snippets to train the time series anomaly detection using Reinforcement learnning Algorithms There is lot be added in this space
TODO:
Have a Generic Time series Dashboard to Annotate anomalies for the Given data
- Use some learning technique to speed up the labelling (manual-labour intensive process)
- Collect the labelled anomaly detection as separate dataset which is able to combine with rest of the features used to train ML models
- Speed up the label function by also using clustering/statistical /distance measure to aid with the Human Oracle to flag an anomaly