alculated within-class scatter matrix is shown below,

ܵௐൌቀ^3.30

െ0.55

െ0.55

6.60 ^ቁ

etween-class scatter matrix can be calculated using the following

where ࢛ was the mean vector of whole data set,

ܵ஻ൌ෍ܰ௞ሺ࢛௞െ࢛ሻሺ࢛௞െ࢛ሻ^௧

௄

௞ୀଵ

(3.13)

wing the above analysis, the estimated mapping vector was,

ܟൌሺ1.76

0.75^ሻ

DA model for the data shown in Table 3.1 is thus shown below,

ݕොൌ1.76 ൈݔଵ൅0.75 ൈݔଶ

ourth column of Table 3.1 shows the values of ݕො using the above

del. Suppose the cutting point (threshold) was 20, the values of ݕො

verted to the prediction class variable Z, which is shown in the

mn of Table 3.1 and is also binary.

king decision using the Bayes rule for a LDA model

s can be made to classify data points into two classes by

g the LDA predictions (projections) against a threshold so that

predictions are converted to the predicted class labels which are

o be able to deliver a robust decision making system, the Bayes

been used in the LDA algorithm and many other discrimination

algorithms.

ose a data point belongs to either the class A or the class B. The

unctions (݂஺ and ݂஻) have been built up for both classes of the

tputs (ݕො஺ and ݕො஻ሻ. Using the Bayes rule for decision-making for