Table 3. Values of rmse and r When Training the Model with Different Sets of Features with the Homogeneous Database^a.

features	rmse (%)	r
fingerprint + additives	2.8 (2.8)	0.59 (0.54)
structural + additives	6.9 (3.9)	0.19 (0.27)
electronic + additives	4.1 (5.4)	0.40 (0.65)
fingerprint + structural + electronic + additives	2.8 (2.7)	0.59 (0.57)

^aValues outside the parentheses correspond to the training set, and those within parentheses correspond to the test set.