Please enable JavaScript.
Coggle requires JavaScript to display documents.
Introduction to Predictive Modeling: From Correlation to Supervised…
Introduction to Predictive Modeling: From Correlation to Supervised Segmentation : :
Identifying Information Attributes
Use variables for Correlation
Entropy
IG (Information Gain)
How much each variable purifies the model.
Variance Reduction
Creates a purer model
Segmenting Data by progressive attribute selection
Tree Induction
Create simple tree
Measure pureness of each subset
Why?
Data corresponds to singular point
Simple visualization
Easy to understand
Descending order to decision variable
Inexpensive
Supervised Segmentation
Target Variable
Tested with Information Attributes
Questions
Why would it be more favorable to use an entropy model over a tree?
How do we better model a tree that does not have a binary class but rather 3, 4, or 5 classes?