Random Forest Classifier

Overview

Random forest is an ensemble machine learning algorithm that constructs multiple decision trees during training and outputs the mode (classification) or mean prediction (regression) of the individual trees. It is robust to overfitting relative to single decision trees, handles high-dimensional features well, and provides feature importance scores. In cancer biology, random forest classifiers have been applied to microbiome data, genomic features, and multi-omic profiles for diagnostic and prognostic classification.

Used by

  • Synthesized in a narrative review of gut-liver axis dysregulation in cholangiocarcinoma: a random forest classifier using eight bacterial genera achieved AUC 0.92–0.99 differentiating cholangiocarcinoma from hepatocellular carcinoma in gut microbiota profiles (Deng T et al., cited); an oral microbiota three-bacterial-biomarker classifier (Rao et al., cited) achieved AUC 0.981 for iCCA vs HCC distinction PMID:25608663
  • Random-forest classifier used for unsupervised clustering of pathway alteration profiles in 109 PDA cases, revealing subtypes where isolated KRAS-pathway alterations (clusters 1 & 2) correlated with poor prognosis. PMID:25855536
  • Random forest classifier trained on methylation and expression features to assign glioma samples to molecular subtypes (LGm/LGr clusters) in the TCGA pan-glioma study PMID:26824661

Notes

  • Feature importance from random forests can identify the most discriminating microbial taxa, genomic alterations, or clinical variables for a given classification task.
  • Model performance should be evaluated on held-out test sets or by cross-validation; AUC values reported from single-cohort studies are subject to overfitting.
  • Applied broadly across cancer genomics (e.g., mutation burden classification, subtype assignment, drug response prediction).

Sources

  • PMID:25608663 — Gut-liver axis review in cholangiocarcinoma (narrative review reporting random forest classifier performance for CCA microbiome diagnostics)

This page was processed by crosslinker on 2026-05-14. - PMID:25855536

This page was processed by wiki-cli on 2026-05-14. - PMID:26824661

This page was processed by wiki-cli on 2026-05-14.