The value of information from a failure prediction model
In this chapter we discuss first the value of information from a failure prediction model. This is necessary since the gathering of information involves costs. The decision-maker will be scrutinized to determine whether returns balance against costs. We have shown in the previous chapter how the sample information from the discriminant analysis can be used in a decision rule. In Section 8.2 the expected opportunity loss attached to this rule is compared with that of a similar rule, but one in which the sample information is not used, and with that of a random, not optimal, decision rule. Finally, in Section 8.3 we discuss the possibilities and restrictions of the use of two different discriminant models by a lending institution to screen credit applicants.
KeywordsDiscriminant Analysis Decision Rule Good Decision Expected Loss Discriminant Model
Unable to display preview. Download preview PDF.