This module takes a date time series and returns: (a) the underlaying linear trend and (b) the times where there is a change in the trend
Project description
ChangePointDetector
This module returns change points in a date time series, using Kalman filters and EVT as described in https://www.robots.ox.ac.uk/~sjrob/Pubs/LeeRoberts_EVT.pdf
- from ChangePointDetector import ChangePointDetector
- Prepare your time series as data plus Panda dates
- Create the necessary Kalman representation by creating a "session" object by calling the ChangePoint class, e.g.:
Session=ChangePointDetector.ChangePointDetectorSession(data,dates).
- 'SeasonalityPeriods' is an optional input, e.g if your data is sequential months, 12 = calendar month seasonality
- 'ForecastPeriods' is another optional input, indicating how many periods to forecast. Default = 3
- Determine the changepoints by running the ChangePointDetectorFunction on your "session", e.g. Results=Session.ChangePointDetectorFunction()
This will return a "Results" object that contains the following:
- ChangePoints. This is a list of 0s and 1s the length of the data, where 1s represent changepoints
- Prediction. This is the Kalman smoothed actuals, plus a 3 period forecast. Note no forecast will be made if there is a changepoint in the last 3 dates
- PredictionVariance. Variance of the smoothed actuals and forecast
- ExtendedDates. These are the original dates plus 3 exta for the forecast (if a forecast has been made)
- Trend. This is the linear change factor
- TrendVariance. Variance of the trend
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Close
Hashes for ChangePointDetectorEVT-0.0.16.tar.gz
Algorithm | Hash digest | |
---|---|---|
SHA256 | 7d15fbe455cdb8767ac38547bf6b5cf52fa85e91c7e223c2302e5507898551a8 |
|
MD5 | 8f13b03774f84782c12fb384d21e2032 |
|
BLAKE2b-256 | ff9baa99e3037c865a03c2d1a7078ff62bbfac0456d2a07f964b611ce34d1931 |
Close
Hashes for ChangePointDetectorEVT-0.0.16-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 356bf03077b824464a41f58c923cc4f013974670f7ce96b5c453c87279301e55 |
|
MD5 | 6ac6b6faaf2e3fb8ab06cf92510e6cb5 |
|
BLAKE2b-256 | df6488a2ef039ec044a79c9f8e4973655e003cba8b16bb0999a437b3269afae1 |