How to import a random forest regression model (made with sklearn (scikit-learn) in python) into Simulink?

Hi, I created a random forest regression model in python using sklearn with the following function:

https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.RandomForestRegressor.html#sklearn.ensemble.RandomForestRegressor

Now, I need to implement this model in Simulink for further use, I tried to save the files with Joblib, ONNX and HDF5 files in python, read these files in MATLAB and then implement it in Simunlink as an LTI-system but with no succes.

Joblib: https://scikit-learn.org/stable/model_persistence.html

used py.pickle.load(py.open('Filename.sav','rb'))

ONNX: https://onnx.ai/sklearn-onnx/

used importONNXNetwork('Filename.onnx')

HDF5: https://calmcode.io/scikit-save/h5.html

https://www.mathworks.com/help/matlab/hdf5-files.html

Any Idea how to import a model created with sklearn in python into Simulink?

Thanks in advance

Answer 1

[Surya]

Hi,

You can make use of MATLAB Python interface. check

You can create a MATLAB function block in Simulink and use python commands directly in the function.

Example MATLAB funtion (For training):

regressor = py.sklearn.ensemble.RandomForestRegressor(n_estimators=100, random_state=0)
regressor.fit(x, y)
py.joblib.dump(regressor, "./random_forest.joblib")

Example MATLAB funtion (For loading the model):

loaded_rf = py.joblib.load("./random_forest.joblib")

For this to work, you need python installed in you system which is compatible with MATLAB. check compatibility

Install required libraries in Python using pip.

Hope it helps.