Roadmap¶
2020 - v0.1 - Initial Feature Set¶
Goals
Provide a set of data structures specialised for time series
Define a standard IO methodss
Provide a set of methods and plots to handle, process and explore time series
dataset.
2020 - v0.2 - Consolidation¶
Goals
This project aims at strengthening loading, preprocessing as well as adding a better visualisation package.
2020 - v0.3 - Modelling, Validation and Detection¶
Goals
This project aims at keeping track of the ongoing models and detectors implementations as well as validation methods. Including :
a set of models allowing for prediction, classification, clustering
and other machine learning tasks on time series * standard methods for time series predictions * a set of methods and plots to validate the quality of a model * a set of models allowing for anomaly detections
2021 - v1.0 - Improve and Publish¶
Goals
This project aims at creating publication and reporting tools to share the results with others. For instance :
Provide a way to publish models e.g. on Kubernetes,
using data from BBData to get an API giving you the prediction,
Allow for retraining of the model from the API.