GetClassificationAccuracyByClassDepth: Classification accuracy at various levels
In thecomeonman/CURD: Implementation of the CURD algorithm for clustering

View source: R/GetClassificationAccuracyByClassDepth.R

At each layer of the tree, calculates the proportion of rows correctly assigned to a particular label. This is calculated across the effective tree if you were to cut all the layers off below the specified layer, i.e. all leaf nodes upto that layer, and all non-leaf nodes in that layer.

1	GetClassificationAccuracyByClassDepth(dtTreeSummary)

dtTreeSummary

Output of TraverseTreeToMeasureClassificationQuality

A summary of the predictions at a particular cut in the tree

Depth NodeMajorityClass NumRows NodeMajorityPc 1: 2 democrat 104 0.9519231 2: 2 republican 128 0.8046875 3: 2 All 232 0.8706897 4: 3 democrat 104 0.9519231 5: 3 republican 128 0.8046875

thecomeonman/CURD documentation built on May 20, 2019, 7:37 a.m.