Confusion: To Similarity...

A command that creates a Similarity from every selected Confusion.

Settings

Normalize
when on, normalize rows by dividing each row element by the row sum. In this way you correct for unequal stimulus numbers.
No symmetrization, Average, Houtgast
determine the symmetrization procedure.

Algorithm

The Average procedure averages:

similarityij= similarityji = (confusionij + confusionji) / 2

The Houtgast procedure as described in the paper by Klein, Plomp & Pols (1970), expresses similarity between stimuli i and j by the number of times that stimulus i and j have resulted in the same response, summated over all response categories.

We use the following formula to calculate the Houtgast dissimilarities:

similarityij = ∑k=1..numberOfColumns min (confusionik, confusionjk)

which is equivalent to the formula in the Klein et al. paper:

similarityij = ∑k=1..numberOfColumns (confusionik + confusionjk – |confusionikconfusionjk|)

Links to this page


© djmw, April 7, 2004