Simple Exponential Smoothing - Heisenberg0203/GETS GitHub Wiki

Simple Exponential Smoothing (SES) applies a weighted average to historical observations to forecast values in future while assuming there is no trend and seasonality in time series.

Forecast equation for SES:

levelt = levelt-1 + (alpha) * (observedt-1 - levelt- 1) 

forecastt + k = levelt 

Grammar for SES:

The smoothing of level in a series is controlled by parameter alpha, which is evolved using the above grammar to minimise the forecast error. The value of alpha and lag is derived through alpha_var and step_var which is then substituted in the above equation. In SES, the value of beta and gamma is substituted as zero, since it assumes there is no trend and seasonality.

Sample Derivation Tree:

Executing SES ( Please make sure that you are in src folder)

python ponyge.py --parameters gets/ses.txt

By default it runs on Daily Waste Generated dataset. Command to run on the custom dataset:

python ponyge.py --parameters gets/ses.txt --dataset_train path_to_traindataset --dataset_test path_to_testdataset.