I have a clustering model which I want to use to detect anamolies in input data. However when I create DMX queries in the SSAS Mining Model Prediction Tab using the PredictCaseLikelyHood function, the function usually returns 0, 1 or NaN without any
apparent correlation to the underlying data. As far as I can tell from the clusters, the underlying data is not particularly random.
Any ideas to what could be causing this?
View Complete Post