A Fisher
Projection was then used to reduce the 200-dimensional data to three
dimensions.
Fisher
Projected Training Data for Asian, Black,
and Hispanic
![](images/fisherRaceTrainNoWhite.png)
[larger
version]
Fisher
Projected Training Data for Asian, Black,
and Hispanic, and White
![](images/fisherRaceTrainWithWhite.png)
[larger
version]
Alternative
View of Complete Training Data
![](images/fisherRaceTrainWithWhiteAlt.png)
[larger
version]
Clearly,
the data is not linearly separable. We will need to use a classification
technique capable of yielding non-linear decision
boundaries.