This script applies a simple transformation to a dataset in order to cast time series forecasting as a supervised learning problem.
The input dataset should include one or more time series as numeric fields. For each numeric field in the input, the script generates additional numeric fields containing row-shifted values of the original field. The user may then use the shifted values as predictors in any supervised learning model.
In addition to the input dataset id, the script takes as input two integers defining the limits of the sliding window, i.e. the minimum and maximum row shifts to consider. Naturally, to implement a forecasting model, the maximum shift should be less than 0 in order to only consider past values as inputs.
Check this readme for an example of the application.
This script applies a simple transformation to a dataset in order to cast time series forecasting as a supervised learning problem.
The input dataset should include one or more time series as numeric fields. For each numeric field in the input, the script generates additional numeric fields containing row-shifted values of the original field. The user may then use the shifted values as predictors in any supervised learning model.
In addition to the input dataset id, the script takes as input two integers defining the limits of the sliding window, i.e. the minimum and maximum row shifts to consider. Naturally, to implement a forecasting model, the maximum shift should be less than 0 in order to only consider past values as inputs.
Check this readme for an example of the application.
This script applies a simple transformation to a dataset in order to cast time series forecasting as a supervised learning problem.
The input dataset should include one or more time series as numeric fields. For each numeric field in the input, the script generates additional numeric fields containing row-shifted values of the original field. The user may then use the shifted values as predictors in any supervised learning model.
In addition to the input dataset id, the script takes as input two integers defining the limits of the sliding window, i.e. the minimum and maximum row shifts to consider. Naturally, to implement a forecasting model, the maximum shift should be less than 0 in order to only consider past values as inputs.
Check this readme for an example of the application.
