Auto Regression - Heisenberg0203/GETS GitHub Wiki

Autoregressive Time Series Forecasting is a type of regression model which is used to predict a variable based on a linear combination of input values.

Forecast equation for Auto_Regression (AR):


X_t = B₀ + B₁* X_t-1 + B₂* X_t-2+ .... + + B_N* X_t-N

Grammar for AR:

This grammar generates values for each of , and . The generates a value which is not a function of time while derives smoothing coefficients for lag variables ranging between -1 to 1. The variable is the input value to the forecaster as a function of time.

Sample Derivation Tree:

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

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

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

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

Please make sure that your dataset has rows more than 365, or else it will derive relation with itself. If the dataset doesn't contain more than 365 instances, please change the lag variable in grammar file to avoid wrong inferences.