Description Usage Arguments Value
View source: R/OptimizeOrdinalFeatureChunks.R
Numeric features can be coarse-classed into bins (e.g. based on quantile limits) so that they resemble ordinal features. However, in the interest of not having too many splits on a node, adjacent bins may be combined in order to arrive at fewer splits. If an ordinal feature has $p$ bins and the maximum number of splits one wishes to have is $q$, then the number of distinct combinations to be tried out is $\sum_r=1^qC(p-1,r-1)$ – this number can be reduced if one can use domain knowledge to rule out some possibilities. The optimal split of an ordinal feature can be determined in the same way that the best features for splitting a node is determined: classifier performance.
1 2 | OptimizeOrdinalFeatureChunks(mValConfMat, nMinSplits = 2,
nMaxSplits = 3)
|
mValConfMat |
todo |
nMinSplits |
todo |
nMaxSplits |
todo |
todo
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.