Table 1 Q² and P² values for amine data set. The mean and standard deviations for the P² and Q² values obtained with the amine data set using two different variable selection methods (corrfilter and PLSfilter) and two different regression methods (PLS and RR). The 5-fold cross validation procedure was repeated 100 times, using 100 different random partitions of the data. N _D and N _L values were selected in an inner 5-fold cross validation loop by optimizing the Q² value. For one random partition of the amine data into five cross validation groups, one P² and five Q² values were obtained. For every random partition the mean Q² is computed. The mean and standard deviations were computed based on the 100 P² values and the 100 mean Q² values.

From: Unbiased descriptor and parameter selection confirms the potential of proteochemometric modelling