Table 3 Variable importance rankings for the best AutoML model algorithm (DNN)
Variables ID | Ranking | Relative importance | Variables |
|---|---|---|---|
GB_txtmean | 1 | 1.000 | Texture mean based on Gabor filter |
GB_R | 2 | 0.935 | R based on Gabor filter |
GMRF6 | 3 | 0.922 | 6th parameter of Gauss Markov random field |
GLCM_corr | 4 | 0.901 | Correlation based on grey-level co-occurrence matrix |
T_line | 5 | 0.897 | Line-likeness based on Tamura algorithm |
GMRF11 | 6 | 0.894 | 11th parameter of Gauss Markov random field |
GMRF12 | 7 | 0.833 | 12th parameter of Gauss Markov random field |
GMRF7 | 8 | 0.818 | 7th parameter of Gauss Markov random field |