Methods to Assess the Predictivity of a QSAR

Although statistical tests can be used to rule out potential structure-activity relationships, they do not guarantee that models that pass such a test will have predictive value. 

One main strategy to test the predictivity of a relationship involves leaving some of the observations out of the modeling process and using the model to predict their potencies. In the leave-one-out (LOO) cross-validation process, each compound is deleted once, a model is developed from the remaining compounds, and the potency of the left-out compound is predicted. The deviation of the observed from this predicted is then used to calculate q, analogous to F . Other cross-validation strategies leave out more of the data on each round of calculation. 

Analogues included in the model Standard error of fit Number of predictions Standard deviation of predictions 

QSAR Alkyl esters plus six more polar 0.13 38 0.429 

---