Entropy dalam machine learning dan AI
Entropy
Assume we have P predictors and K classes. Suppose we select the kth predictor and split a region along the threshold .
We can assess the quality of this split by measuring the entropy of the class distribution in each newly created region by calculating:
We can now try to find the predictor p and the threshold tp minimizes the weighted average entropy over the two regions:
Where Nr is the number of training points inside of region Rr.
Comments
Post a Comment