siso_predictor

A predictor of future behavior of a targeted criterion based on input/output profiles

Code Citation

Possible use

• Forcasting future evolution of an uncontrolled variable from its past
• Design a Data-Driven Nonlinear Model Predicive Control

Required Packages

numpy, matplotlib

Contents

• The siso_predictor module siso_predictor.py
• The main.py file
• A utility module generate_data.py that is used to generate the data for the test of the module

Example of use

sol = learn_model(
    y=y,
    U=U,
    ydef=ydef,
    N=100,
    n_clusters=3,
    nJump=1,
    max_leaf_nodes=1200,
    test_size=0.33,
    validation_mode='all',
    plot=True
    )

where

• y the output time series
• U the input time series
• ydef the map that defines the target indicator to predict
• N the window width
• n_clusters the number of cluster used in the predictor
• nJump the jump size used when processing the data
• max_leaf_nodes the maximum number of leaf nodes in the Random Forest predictor
• test_size the test size used in the learning validation split of the data
• validation_mode the visualisation mode of the result ('all', 'learning', 'test')
• plot whether to plot the results or not.

Example of result