1. Title: Hill-Valley Dataset 

2. Source Information

   a)	Creators:

	Lee Graham (lee@stellaralchemy.com)

	Franz Oppacher (oppacher@scs.carleton.ca)
	Carleton University, Department of Computer Science
	Intelligent Systems Research Unit
	1125 Colonel By Drive, Ottawa, Ontario, Canada, K1S5B6

   c) Date of release: March 2008
 
3. Past Usage:

   (a) Non-published. Evaluation of dataset by various learning algorithms in the Waikato Environment for Knowledge Analysis (WEKA).


4. Relevant Information:

   Each record represents 100 points on a two-dimensional graph. When plotted in order (from 1 through 100) as the Y co-ordinate, the points will create either a Hill (a bump in the terrain) or a Valley (a dip in the terrain).

   There are six files, as follows:

   (a) Hill_Valley_without_noise_Training.data
   (b) Hill_Valley_without_noise_Testing.data

   These first two datasets (without noise) are a training/testing set pair where the hills or valleys have a smooth transition.

   (c) Hill_Valley_with_noise_Training.data
   (d) Hill_Valley_with_noise_Testing.data

   These next two datasets (with noise) are a training/testing set pair where the terrain is uneven, and the hill or valley is not as obvious when viewed closely. 

   (e) Hill_Valley_sample_arff.text

   The sample ARFF file is useful for setting up experiments, but is not necessary.

   (f) Hill_Valley_visual_examples.jpg

   This graphic file shows two example instances from the data.


5. Number of Instances: 606 for each training and testing set

6. Number of Attributes: 100 predictive attributes, 1 goal attribute

7. Attribute Information:
   1-100: Labeled X##. Floating point values (numeric)
   101: Labeled class. Binary {0, 1} representing {valley, hill}

8. Missing Attribute Values: None

   There is no class noise. The noisy datasets are named as such because it more accurately represents the terrain.

9. Class Distribution:

   Hill_Valley_with_noise_Training.data (307 / 299)
   Hill_Valley_with_noise_Testing.data  (299 / 307)

   Hill_Valley_without_noise_Training.data (305 / 301)
   Hill_Valley_without_noise_Testing.data  (295 / 311)
