Predictor variables essential to accurate land cover classification were selected using importance indices statistically derived with classification trees.
Four solids (two Ballotini, a silica sand and a spent FCC), belonging to groups A and B of Geldart's classification, were selected in order to separately investigate the effect of particle size and density on axial profiles of voidage and on radial profiles of solids mass flux.
The potential variables contributing to the classification were selected based on variable importance in the projection (VIP > 2.0) values.
After this processing, the most significant features for the classification were selected, and the training and testing set were created with a function of probability distribution similar to the one that provided the best results in the classification.
Using random Forest supervised classification algorithm, nine lncRNAs mostly related to the prognostic classification were selected among the 909 lncRNAs (figure 1C) according to the permutation important score by the software Random Jungle (see online supplementary methods).
In the first step, all univariate models were tested and only probes significantly associated with the MSP classification were selected using the log-likelihood ratio test (LRT) and Bonferroni correction for multiple testing.
FOXP3+ T-cell density of the SLN, primary tumour size, and Lauren classification were selected as significant predictors of non-SLN metastasis, including micrometastases and macrometastases, by this approach.
The factors included in the modified PVN classification were selected based on the precise analyses of this study using well-known clinicopathological factors, such as histologic grade, invasive tumour size, and nodal status.
