Guided Random Forest in the RRF Package
Houtao Deng
Abstract
Open-access reader
Houtao Deng
Abstract
Open-access reader
Random Forest (RF) is a powerful supervised learner and has been popularly used in many applications such as bioinformatics. In this work we propose the guided random forest (GRF) for feature selection. Similar to a feature selection method called guided regularized random forest (GRRF), GRF is built using the importance scores from an ordinary RF. However, the trees in GRRF are built sequentially, are highly correlated and do not allow for parallel computing, while the trees in GRF are built independently and can be implemented in parallel. Experiments on 10 high-dimensional gene data sets show that, with a fixed parameter value (without tuning the parameter), RF applied to features selected by GRF outperforms RF applied to all features on 9 data sets and 7 of them have significant differences at the 0.05 level. Therefore, both accuracy and interpretability are significantly improved. GRF selects more features than GRRF, however, leads to better classification accuracy. Note in this work the guided random forest is guided by the importance scores from an ordinary random forest, however, it can also be guided by other methods such as human insights (by specifying $λ_i$). GRF can be used in "RRF" v1.4 (and later versions), a package that also includes the regularized random forest methods.
OpenAlex reports 61 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Random Forest (RF) is a powerful supervised learner and has been popularly used in many applications such as bioinformatics. In this work we propose the guided random forest (GRF) for feature selection. Similar to a feature selection method called guided regularized random forest (GRRF), GRF is built using the importance scores from an ordinary RF. However, the trees in GRRF are built sequentially, are highly correlated and do not allow for parallel computing, while the trees in GRF are built independently and can be implemented in parallel. Experiments on 10 high-dimensional gene data sets show that, with a fixed parameter value (without tuning the parameter), RF applied to features selected by GRF outperforms RF applied to all features on 9 data sets and 7 of them have significant differences at the 0.05 level. Therefore, both accuracy and interpretability are significantly improved. GRF selects more features than GRRF, however, leads to better classification accuracy. Note in this work the guided random forest is guided by the importance scores from an ordinary random forest, however, it can also be guided by other methods such as human insights (by specifying $λ_i$). GRF can be used in "RRF" v1.4 (and later versions), a package that also includes the regularized random forest methods.
Key concepts: Random forest, Interpretability, Feature selection, Feature (linguistics), Computer science, Selection (genetic algorithm), Random testing, Random effects model