Waveform dataset
The information provided here has been largely replicated from the notes
accompanying the waveform dataset in the UCI repository of
machine learning databases.
1. Title: Waveform Database Generator (written in C)
2. Source:
- Breiman,L., Friedman,J.H., Olshen,R.A., & Stone,C.J. (1984).
Classification and Regression Trees. Wadsworth International
Group: Belmont, California. (see pages 43-49).
- Donor: David Aha
Date: 11/10/1988
3. Past Usage:
CART book (above):
- Optimal Bayes classification rate: 86% accuracy
- CART decision tree algorithm: 72%
- Nearest Neighbor Algorithm: 38%
- 300 training and 5000 test instances
4. Relevant Information:
- 3 classes of waves
- 21 attributes, all of which include noise.
- See the book for details (pp 49-55, 169)
5. Attribute Information:
- Each class is generated from a combination of 2 of 3 "base" waves
- Each instance is generated with added noise (mean 0, variance 1) in
each attribute
- See the book for details (49-55, 169)
6. Missing Attribute Values:
none
7. Class Distribution:
33% for each of 3 classes