4 ms·
They tend to create difficult to interpret models that don't perform as well as other "black box" modeling methods (GBMs, neural nets, etc.)
by rcar 9y ago
They tend to create difficult to interpret models that don't perform as well as other "black box" modeling methods (GBMs, neural nets, etc.)
- closed 9y agoWas this true, or perceived as true in 2003? My understanding was that people did not see them as performing worse than NN back then.
- rm999 9y agoDefinitely true. I worked for a company that was generating millions of dollars a year from neural networks in the mid 90s (edit: to be clear I didn't work there in the 90s, I joined years after their initial buildouts). The Unreasonable Effectiveness™ of neural networks has been true for a long time. When I worked there I tried switching out some models with SVMs and they were less accurate and took 1-2 orders of magnitude more time to train.
- closed 9y agoReally useful to hear, thanks! I know psychology was gaining a lot of headway with NN models in the 90s, but had little sense for what was going on in industry.
- abhgh 9y agoThis is not really true. Aside from ensemble models they tend to perform pretty much at par or better. Here's an extensive comparison by Rich Caruana [1] [1] https://www.cs.cornell.edu/~caruana/ctp/ct.papers/caruana.icml06.pdf https://www.cs.cornell.edu/~caruana/ctp/ct.papers/caruana.ic...