Future forecasting, log(1+x), and a MovingAverage transformer

EthanRosenthal / skits

scikit-learn-inspired time series

https://www.ethanrosenthal.com/2018/03/22/time-series-for-scikit-learn-people-part2/

MIT License

197 stars 17 forks source link

Future forecasting, log(1+x), and a MovingAverage transformer #13

Closed cbartondock closed 5 years ago

cbartondock commented 5 years ago

I added an end_idx to the forecast method so that if desired the user can forecast further into the future than they have data for. IMO Log transform should always be log(1+x) so that zero doesn't give -Inf. Completed the MA part of ARIMA :)