Skip to main content

Table 9 Top five important descriptors selected by three blending methods

From: Comparison and improvement of the predictability and interpretability with ensemble learning models in QSPR applications

Uniform blendingLinear blending
Selected descriptorsFeature importanceSelected descriptorsFeature importance
HeavyAtomCount0.05993HeavyAtomCount0.04781
NumRotatableBonds0.04803NumRotatableBonds0.04219
wing2_HeavyAtomCount0.03827wing2_HeavyAtomCount0.03471
mesogen_HeavyAtomCount0.03484wing1_HeavyAtomCount0.03006
wing1_HeavyAtomCount0.03290fr_unbrch_alkane0.02897
Any blending  
Selected descriptorsFeature importance  
HeavyAtomCount0.06520  
NumRotatableBonds0.05236  
wing2_HeavyAtomCount0.04321  
wing1_HeavyAtomCount0.03636  
mesogen_HeavyAtomCount0.03556