Comment

Comment by Miles Garnsey on July 6, 2011 at 1:55am

Perhaps because they have familiarity with a wider range of models than other analysts... Which means they have an understanding that other groups might not of different strengths and weaknesses of models, having used different models from many different disciplines.

On the other hand, an analyst in corporate or academic spheres is more likely to use only models which are conventional in their own field, and thus have less knowledge of other competing models and their strengths.

Comment by Jimmy Wong on July 5, 2011 at 11:53am

Why would data mining consultants be more likely than the other groups to use Ensemble Models?  Do you think it's more because of the complexity of combining the multiple models sufficiently well in-house, or simply the in-house effort required outweighs the benefits?  The chart and article described the observations of Ensemble Models usage but did not go into details as to reasons why, or how it would be significant.