Visual Decision Tree (2007)
1) Data format
|------------------------------------------------
| File Description
|------------------------------------------------
| Classes = (ClassName n)[, ...]
| Attributes = (Attributes {Type | Categorical Values})[, ...]
| value nm [, value nm+1, ...], Class Index
| value (n+1)m [, value (n+1)m+1, ...], Class Index
Classes = Chemise, Shirt, Kaki Trousers, Jean Trousers
Attributes = Size{Integer}, Cost{Floating_Point},
Cloth{Floating_Point}, Weight{Floating_Point}
33, 45.9, 32.4, 0.15, Chemise
33, 48.0, 38.5, 0.17, Shirt
30, 78.8, 94.1, 0.34, Kaki Trousers
28, 67.6, 56.8, 0.29, Jean Trousers
26, 64.7, 86.8, 0.26, Kaki Trousers
2) Data selection
- Open datasets
- Choose visualization methods
- Left mouse for bushing
- Right mouse for selection
- Tools menu
- File menu for data extraction
3) Visual Tree Builder
a) Tree construction
- Open the dataset
- Choose visualization methods
- Left mouse for splitting design
- Right mouse for splitting
- Save the tree model with Model menu
- Others: Backtracking, Splitting datasets for learning & testing
b) Classification
- Open a dataset
- Load a tree model to predict
- Result (see node info)
Last update feb 18 2007 by Thanh-Nghi Do
CS 83818 - 29238 Brest Cedex 3 - France
Tel: +33 (0)2 29 00 11 11
Email: tn.do-AT-telecom-bretagne-DOT-eu