Equal-width Interval BinningΒΆ
Global Algorithm - One-Dimensional Algorithm
Equal-width Interval Binning algorithm divides the range of values into \(k\) subintervals of equal width that is determined by
\[h = \frac{max(Y)-min(Y)}{k}.\]
Input Parameters
Parameter | Type | Constraint | Description | Remarks |
---|---|---|---|---|
\(Y\) | \(Y \in \mathbb R^N\) or \(Y\) is continuous data | For discrete data, \(N \in \mathbb{N}\) ; for continuous data, input data bounded from below and above | ||
\(k\) | \(k \in \mathbb N\) |
Output Parameters
Parameter | Type | Constraint | Description | Remarks |
---|---|---|---|---|
\(\hat{Y}\) | \(\hat{Y} \in \mathbb R^{k+1}\) | Discrete data set of length \(k+1\) |
Single Steps using the Algorithm
- Data Discretization with Equal-width Interval Binning
- Data Reduction With Equal-width Interval Binning
References
J. Dougherty, R. Kohavi, M. Sahami, Supervised and Unsupervised Discrimination of Continuous Features, Proceedings of the 12th International Conference, Morgan Kaufman, pp. 194-202, 1995.
http://robotics.stanford.edu/users/sahami/papers-dir/disc.pdf