Description:
Classification model developed to predict the Percentage of Repellency, PR, (%) in three breeds of cockroach (Blatella germanica, Periplaneta americana, and Blatta orientalis) in two classes: ACTIVE or INACTIVE. The breakpoint is 90 %. Values greater than or equal to the breakpoint will elicit a repellent response in these specific cockroaches and are represented as ACTIVE. Lower values represent certain actions occurring, however, these are not enough to activate the repellent response, these are classified as INACTIVE.

The training was performed with the Stacking algorithm as default in Weka 3.9.4 with 10-fold cross-validation by using the Quadratic Discriminant Analysis function as meta classifier and as classifiers: IBk (with K-nearest neighbors = 10 and True cross-validation), John Platt's sequential minimal optimization algorithm for training a support vector classifier (SMO - with Pearson Universal Kernel (PUK)), and Random Forest. A number of 2 QuBiLS-MIDAS descriptors are in the classification model. The QuBiLS-MIDAS descriptors are namely:

AC[5]_K_Tr_AB_nCi_3_M21(M8)_SS1_T_KA_psa-a-c_MID
GV[6]_K_TrB_AB_nCi_3_M22(M15)_SS1_A_KA_r-p_MID

Training set:
34 compounds extracted from 10.1002/cbdv.200890058

Performance:
For a 10-fold cross-validation, the statistical parameters (Performance without applicability domain) are MCC = 1, ROC Area = 1, PRC Area = 1, TP Rate = 1, FP Rate = 0, Q (%) = 100, and Precision = 1.

Reference:
Gaudin et al. Carboxamides Combining Favorable Olfactory Properties with Insect Repellency. 2008, 5(4), 617-635. DOI: 10.1002/cbdv.200890058