Skip to main content
Fig. 3 | BMC Bioinformatics

Fig. 3

From: High dimensional model representation of log-likelihood ratio: binary classification with expression data

Fig. 3

The overall pipeline of LAS-HDMR and LABS-HDMR. In step 1 the algorithm parameters and the internal classification machineries are selected. In step 2 classifiers using individual features and feature pairs are trained and log likelihood ratios are saved. In step 3 weak classifiers (based on rf and \(|{r}^{{f}_{i},{f}_{j}}|\)) are removed. LAS-HDMR moves to step 5, but LABS-HDMR moves to steps 4a and 4b, in which feature pairs construct blocks, and weak blocks are removed. In step 5 the final feature vector V(X) is constructed and in step 6 the weights used to combine elements of V(X) are computed

Back to article page