ChangePointDetectorEVT 0.0.18

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

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

1. from ChangePointDetector import ChangePointDetector
2. Prepare your time series as data plus Panda dates
3. 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
4. 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