4 ms·
It's pretty easy to visualize decision trees and there are already open source usable tools for that. It would be way cooler if they could visualize Random For
by lqdc13 10y ago
It's pretty easy to visualize decision trees and there are already open source usable tools for that.
It would be way cooler if they could visualize Random Forest that can have thousands of these trees.
This is what is actually used more frequently in practice.
- huac 10y agoYeah, decision trees really shouldn't be used in practice. The variance of any single decision tree is too high to give you reasonable results. See also http://scott.fortmann-roe.com/docs/BiasVariance.html http://scott.fortmann-roe.com/docs/BiasVariance.html
- binalpatel 10y agoThey have their place - if you're more interested in inference than in prediction they can be useful tools.
- huac 10y agoDo you have a source for this? Pretty curious to see how that would play out.
- binalpatel 10y agoNo formal sources unfortunately, it's similar to using a linear regression model trained on historical data for inference (with some cross-validation/penalization to make sure we're not just finding signals in the noise). We know for sure that there's a ton of bias (in the bias-variance sense), and that the predictions won't be nearly as good as a blackbox model, but we may glean some useful insights out of it. Same idea with decision trees, in the past I've used them to find what factors influence customer support scores on online tickets, and we found some surprising and useful insights that were used for process improvements.
- papaf 10y agoYou don't need a source -- just read or print a decision tree after training one.