Class Prediction with Linear Discriminant Analysis(LDA)

In this project we tried to predict the class of a given point in a plane where we randomly add points in two different colors.

Video

How it works?

Creating visualizing the training set:

• After you provide limit points by given X and Y position of it, app will randomly distribute the points in two different colors on plane. You can use Limit X and Limit Y inputs.
• You should provide a number of points for the training set for its prediction. App will generate the points as much as this value.
• Blue points represents the class 1. Reds represent the class 2.
• When you hit the Start button, you areoing to see the plane with the dots in two different color(blue-red).

Adding a test point, for prediction the class of it:

• Once you generate the plane you can now add a test point by providing the X and Y position of it. You can use Test X and Test Y inputs.
• After that hit Add button. Then you should see the green point on the plance.
• Sometimes it is not showing on the plane because of the other points density. You in this case please see the output text window to confirm its add.
• Once you hit add button Class should be value should set to what class should be after the prediction. You can use this information as a comparision value with the prediction.

Starting the analysis:

• Click Analysis button for seeing the results of the prediction.
• This willive you two different prediction result. Discriminant function predicted class and Mahalanobis distance predicted class

You can add more test point without reseting the training set and the plane. If you want to start over use Reset button.

Test

Limit X	Limit Y	Num.Points	Test X	Test Y	Actual Class	Disc. Pred. Class	Mah. Pred. Class
300	300	1000	10	10	1	1(True)	1(True)
300	300	1000	250	250	1	2(False)	1(True)
300	300	1000	280	250	1	2(False)	2(False)
300	300	1000	400	250	2	2(True)	2(True)