LexGraphTutor

# LexGraphTutor: Structural Feature Modeling and Classification of Dramatic Texts for Legal Education This repository contains research and code implementations from the paper **"Research on Structural Feature Modeling and Classification Methods of Dramatic Texts for Legal Education"** by Yanlong Jin. It presents a structured framework for improving legal education through dramatic texts, integrating **structural modeling, narrative classification, and adaptive instruction**. --- ## Features - **LexGraph**: A hierarchical graph-based model representing legal knowledge, procedural transitions, and jurisdiction-sensitive branching. - Models legal logic flows and interpretative layers - Supports jurisdiction-specific curriculum projections - Captures temporal evolution of legal doctrines - **LexTutor**: An adaptive instructional strategy leveraging LexGraph to generate **personalized learning pathways**. - Learner profile–based customization - Constraint-driven instructional planning - Context-aware adaptive legal instruction - **Hybrid NLP Models**: Combines **dialogue-aware transformers** and **rule-based semantic modules** for better interpretability and classification of dramatic legal texts. --- ## Results - Achieved **12% improvement in classification accuracy** over baseline models (BERT, RoBERTa, DeBERTa, etc.) across multiple legal NLP datasets: - **JRC-Acquis** (Multilingual Legislative Texts) - **ContractNLI** (Contract Inference Tasks) - **CLaF** (Cross-lingual Legal Analysis) - **LEVEN** (Legal Event Extraction) - Demonstrated improved scalability and interpretability in legal pedagogy applications, enabling **interactive legal instruction environments**. (See *Figures 1–4* in the paper for detailed benchmarking.) --- ## Framework Overview ### LexGraph- Jurisdiction-aware curriculum modeling- Doctrinal evolution encoding- Normative dependency representation ### LexTutor- Constraint-based path planning - Adaptive value-driven selection - Computationally efficient planning (O(T·|N|²)) (*Figures 5–8* illustrate the architecture.) --- ## Installation ```bashgit clone https://github.com/your-username/lexgraphtutor.gitcd lexgraphtutorpip install -r requirements.txt