212
8: Stefan A. Robila, Pramod K. Varshney
a
e
g
Fig.8.7a-j. Components produced by the rCA-FE algorithm. a soybean; b wheat and grass
pasture; c roads and towers; d corn; e trees; g pasture
to differentiate between the classes wheat and the grass pasture, which was not
the case with leA-FE. The wheat was projected in dark, in the lower side of the
Fig. 8.8f, while the grass pasture has been projected in bright in the left side of
the Fig. 8.8c.
We employ mutual information as a quantitative measure of how well the
classes have been separated among various components. For an m dimensional
8: Stefan A. Robila, Pramod K. Varshney
a
e
g
Fig.8.7a-j. Components produced by the rCA-FE algorithm. a soybean; b wheat and grass
pasture; c roads and towers; d corn; e trees; g pasture
to differentiate between the classes wheat and the grass pasture, which was not
the case with leA-FE. The wheat was projected in dark, in the lower side of the
Fig. 8.8f, while the grass pasture has been projected in bright in the left side of
the Fig. 8.8c.
We employ mutual information as a quantitative measure of how well the
classes have been separated among various components. For an m dimensional