Bhattacharyya Distance Measure for Pattern Recognition
by Yi Cao
08 Feb 2008
(Updated 13 Mar 2008)
A Tool to calculate the Bhattacharyya Distance Measure for Pattern Recognition
|
Watch this File
|
| File Information |
| Description |
The m-file provides a tool to calculate the Bhattacharyya Distance Measure (BDM) between two classes of normal distributed data. The BDM is widely used in Pattern Recognition as a criterion for Feature Selection.
Directly calculation may result in divide by zero error due to possible (near) singularity of cov(X1)*cov(X2). The improved code uses the Cholesky factorization for the normal cases but uses sqrtm for near singular cases. |
| MATLAB release |
MATLAB 7.5 (R2007b)
|
|
Tags for This File
|
| Everyone's Tags |
|
| Tags I've Applied |
|
| Add New Tags |
Please login to tag files.
|
| Updates |
| 11 Feb 2008 |
update example and improve speed |
| 12 Feb 2008 |
update code to include numerical reliability check. |
| 13 Mar 2008 |
update descriptions. |
|
Contact us at files@mathworks.com
