文章目录
-
-
-
-
- decision tree induction
- General structure of Hunt's algorithm
- Design issues of decision tree induction
- Methods for expressing test conditions
- Test condition for nominal attributes
- Splitting based on continuous attributes
- How to determine the best split
- Alternative Splitting Critera beased on INFO
- Split Based on INFO
- Finding the best split
- Measure of impurity: GINI
- Computing Gini Index for a collection of nodes
- Binary attributes: computing GINI Index
- Categotical attributes: computing Gini Index
- Measure of Impurity: Entropy
- Computing Information Gain after splitting
- Problems with Information Gain
- Gain Ratio
- Measure of Impurity: Classification Error
-
-
-
illustrating classification task
classification techniques
-
example of a decision tree
-
apply model to test data
-
decision tree classification task
decision tree induction
-
Hunt’s Algorithm(one of the earliset)
multi-way split or binary split
How to determine the best split
greedy approach: nodes with homogeneous class distribution are preferred
G i n