CLASSIFICATION
- Created by: smrc
- Created on: 21-08-19 18:04
Fullscreen
To find a clasifier that can map individual data items into one of several pre-defined and non over-lapping classes.
- SUPERVISED learning method: Labelled records, training data used to define classes.
Classification Tasks:
- Mortgage
- Credit Risk
- Heart Disease
Historical data = training data.
ID3 + data = Decision Tree + Defined Classes
Values of attributes of unseen data compared to values in decision tree, path created to class.
Classification is non-overlapping (usually binary)
DECISION TREE: A classification method used either to calculate probability that a given…
Comments
No comments have yet been made