Abstract
Presented herein is a novel algorithm for inference on decision forest models that increases the robustness of the decisions in the presence of missing features in the data. The proposed algorithm ensures that tree decisions are supported by a minimal amount of non-missing features. Experiments have demonstrated that the proposed algorithm not only increases the robustness of the model, but also increase the model’s predictive performance.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Brabec, Jan and Skarda, Cenek, "ROBUST DECISION FOREST INFERENCE ALGORITHM IN PRESENCE OF MISSING FEATURE VALUES", Technical Disclosure Commons, (January 30, 2020)
https://www.tdcommons.org/dpubs_series/2909