MATLAB: Using feature extraction methods and a classification tree

Question

I saw this page (below) which mentioned the RICA and SparseFilt methods, and I would like to use these methods before classifying data using a binary decision tree.

https://www.mathworks.com/help/stats/feature-extraction.html

How can I accomplish this workflow using built-in functions?

Answer 1

A detailed overview of this workflow from end-to-end can be found at the following link:

https://www.mathworks.com/help/stats/workflow-for-feature-extraction.html

1. The first step here is to calculate the transform model using either the 'rica' or 'sparsefilt' functions. Examples below:

>> myModel = rica( myData , newDimension );
or
>> myModel = sparsefilt( myData , newDimension );

2. After this, you can transform your data into the new space:

>> transformedData = myModel.transform( myData );

For more information (and examples) about using these two functions please refer to the following links:

https://www.mathworks.com/help/stats/rica.html

https://www.mathworks.com/help/stats/sparsefilt.html

3. Now that your data is ready for training,  you can train a Classification Tree using the 'fitctree' function:

>> myTree = fitctree( transformedData , labels );

4. Finally you are ready to create predictions on some other (transformed) dataset:

>> predicatedClasses = myTree.predict( myModel.transform(myOtherData) );

For more information on 'fitctree', you can refer to the following link:

https://www.mathworks.com/help/stats/fitctree.html

Note that 'fitctree' has very many options (link below) and parameters that can be set, as seen below:

https://www.mathworks.com/help/stats/fitctree.html#namevaluepairs