erent in Figure 3.3(b). The proportion of the cross class data

as about 90.9%. Therefore, two posterior probabilities derived

LDA model constructed for this data set were very different, i.e.,

ൌ0.09 and ߨ஼௥௢௦௦ൌ0.91. The shapes and the heights of two

functions shown in Figure 3.4(b) therefore demonstrated a

nt difference.

(a) (b)

The density of estimated based on the LDA predictions (ݕො஺) for data shown in

a). (b) The density of estimated based on the LDA predictions (ݕො஺) for the data

igure 3.3(b).

(a) (b)

he posterior probabilities estimated by LDA for two data sets shown in Figure

lid line stands for the posterior probability curve for the triangle class and the

stands for the posterior probability curve for the cross class. (a) The posterior

curves for the data shown in Figure 3.3(a). (b) The posterior probability curves

shown in Figure 3.3(b).