Using prediction model inside a matlab function block

This is the content in the prediction file.

data1new=readtable('processed_data.csv');

X1 = data1new{:,1:end-1};

Y1 = data1new{:,end};

% Train a Random Forest ensemble using all the data

ensnew = fitrensemble(X1,Y1,'Method','bag','NumLearningCycles',150,'Learners','tree');

NH5pred=predict(ensnew,X1);

%saving the model for future use.

save('NH5predModel.mat','ensnew');

The image shows the data being collected and the MUX block delivers the input to the function block.

Inside the matlab function block:

function y = Regpred(inputt)

% Load the regression model

loadedData = load('NH5predModel.mat', 'ensnew');

ensnewLoaded = loadedData.ensnew;

% Ensure input has 8 columns

if size(inputt, 2) ~= 8

error('Input must have 8 columns.');

end

% Preallocate the output variable y based on the size of input

numRows = size(inputt, 1);

y = zeros(numRows, 1); % Assuming a single output value per row of input

% Make predictions using the loaded model

for i = 1:numRows

y(i) = predict(ensnewLoaded, inputt(i, :));

end

The outport the matlab function block is a To workspace block.

The error encountered is also attached. Please help!

Answer 1

[Drew]

MathWorks provides Simulink blocks for machine learning model prediction. So, rather than using a generic MATLAB function block for the model prediction, you can use the native Simulink block designed for model prediction for your random forest model. Check the "Blocks" section of the doc page https://www.mathworks.com/help/stats/code-generation.html.

Given that your model is a random forest trained with fitrensemble, the matching Simulink block is the RegressionEnsemble Predict block:

• RegressionEnsemble Predict Predict responses using ensemble of decision trees for regression (Since R2021a)

The sidebar indicates that you are using R2020a. Note that this "RegressionEnsemble Predict" block was added in R2021a.

If this answer helps you, please remember to accept the answer.

[Prana]

Thank you Mr Drew. Will surely try that.