MORF: Multi-view oblique random forest for hepatotoxicity prediction
Binsheng Sui1, Qingzhuo He2, Bowei Yan3
1Department of Digital Media, Xiamen University, Xiamen 361005, China.
Iscience
|February 5, 2025
Summary
This study introduces a novel Multi-View Oblique Random Forest (MORF) for accurate hepatotoxicity prediction. The MORF effectively utilizes diverse feature types, enhancing drug development reliability.
Area of Science:
- Computational chemistry and toxicology
- Machine learning in drug discovery
Background:
- Hepatotoxicity prediction is critical for safe drug development.
- Existing methods may not fully leverage multi-type feature data.
Purpose of the Study:
- To propose a novel Multi-View Oblique Random Forest (MORF) for enhanced hepatotoxicity prediction.
- To address the challenge of integrating diverse feature types in toxicity assessment.
Main Methods:
- Developed a Multi-View Oblique Random Forest (MORF) model.
- Utilized Householder transformation for inclined cut hyperplanes within each feature view.
- Designed two Multi-View Oblique Decision Tree (ODT) algorithms (ODT-N, ODT-R) as base learners.
Main Results:
- MORF algorithms demonstrated effective utilization of information from different feature views.
- Experimental comparisons validated the performance of the proposed MORF approach.
Conclusions:
- The proposed MORF provides an effective and reliable method for hepatotoxicity prediction.
- This approach enhances the integration of multi-view data in toxicological assessments.
