pharo-ai / gaussian-mixture-model Goto Github PK

We should have a method to predict new instances. Not only classifying the ones with which the algo was trained.

The current method labels only returns the labels for the data with the algo what fitted.

testLabels

	model numberOfGaussians: 2.
	model fitX: #( #( 1 2 ) #( 1 4 ) #( 1 0 ) #( 10 2 ) #( 10 4 ) #( 10 0 ) ).

	self
		assert: model labels
		equals: #( 1 1 1 0 0 0 )

This test is failing, it returns #( 1 1 1 1 1 1) instead

Now, we depend on master which can give problems if the api change. Which it does

We should have one method is is called hasToStopIterating that check is the algo has converged and also if the max iteration was reached. In that way we can check if the model has converged

None of the classes have comments

AIGaussianMixtureModelTest
✗ #testLabels (56ms)
✗ #testPredictNewData (3ms)
✗ #testPredictSameFittingData (1ms)

Write a wiki page of the algporithm explaining how it works, how to use it and some examples