A new method for linear feature selection is described which has as its underlying theme the preservation of actual distances between training data points in the lower dimensional space. Comparison with existing methodology places the method closer to the principle components or Karhunen- Loève approach than to methods based on an approach through statistical pattern recognition. A computer program implementing the technique is described. An example application to 12 dimensional LANDSAT data is given. © 1979.
Bryant, J., & L.F. Jr., G. (1979). Distance preserving linear feature selection. Pattern Recognition, 11(5–6), 347–352. https://doi.org/10.1016/0031-3203(79)90046-3