A Novel Rotational Causal Random Forest Approach for Word Sense Disambiguation of English Modal Verbs

Abstract

Modal verbs play a crucial role in the context and exhibit significant semantic ambiguity, necessitating effective disambiguation techniques for improved natural language understanding. To enhance the accuracy of word sense disambiguation (WSD) in English modal verbs, this manuscript proposed Rotation Casual Random Forest (RCRF) which integrates rotational causal ratios with decision trees and random forests. To verify the effectiveness and conciseness of the approach, we evaluated it with 10 public dataset of modal verbs and compared it with other 8 existing approaches. The results showed that RCRF maintains a more concise structure compared to other approaches with higher classification accuracy. RCRF achieves a balance between accuracy and efficiency, offering a promising tool for semantic analysis in computational linguistics. This method provides a robust framework for addressing polysemy in modal verbs and has implications for broader applications in natural language processing.

Keywords: Rotation Casual Random Forest, Word Sense Disambiguation, English Modal Verbs.