Logistic Regression Models

This type of model is used when the output parameter is not a continuous variable but takes an absolute value of 0 or 1. The relationship between input and output variables need not be linear in this case. The format of the model usually takes the form of:

All these models are simply attempts to associate or pattern a set outcome with specific inputs^

Once a model is chosen, the question is how good is it?

The receiver operating characteristic curve is the graphical representation of the True Positive Rate (sensitivity) versus the False Positive Rate (l-specificity). The ideal model would have an receiver operating characteristic of l.

TABLE 1 ■ The Meaning of Receiver Operating Characteristic Values

How good is

Area under ROC

the model?

1

Perfect

0.9-1

Excellent

0.8-0.9

Good

0.7-0.8

Fair

0.6-0.7

Average to poor

0.5-0.6

Not of value

Abbreviation: ROC, receiver operating

characteristic.

Duration of hospital stay plays a critical role in determining the cost of individual surgical procedures.

0 0

Post a comment