中医智能辨证辨病与处方决策系统研究——基于多任务BERT模型的一体化实现

Research on Intelligent Syndrome Differentiation, Disease Identification and Prescription Decision System for Traditional Chinese Medicine —An Integrated Implementation Based on Multi-Task BERT Model

摘要: 随着人工智能与中医药的深度融合，中医辨证论治的数字化、智能化已成为推动中医药现代化的重要方向。本研究基于脱敏中医病历数据集，构建了集多标签辨证辨病与中药处方推荐于一体的智能决策系统，实现了从患者临床文本信息到证型识别、疾病诊断及个性化处方生成的一体化处理。系统采用预训练语言模型(BERT)与多任务学习架构，同步完成辨证、辨病与处方推荐三大核心任务。实验以1500条真实脱敏病历为支撑，按8:2:5划分训练、验证与测试集，通过多维度指标评估模型性能。结果表明，系统在辨证辨病任务上的综合准确率达0.52，处方推荐任务的综合评分达0.48，整体性能优于传统单一任务模型与基线模型。本研究为中医智能诊疗提供了可参考的技术方案，有助于提升基层诊疗效率与规范化水平，推动中医药事业的现代化发展。

Abstract: With the in-depth integration of artificial intelligence and traditional Chinese medicine (TCM), the digitalization and intelligentization of TCM syndrome differentiation and treatment have become an important direction for promoting the modernization of TCM. Based on a desensitized TCM medical record dataset, this study constructs an intelligent decision-making system integrating multi-label syndrome differentiation, disease identification, and Chinese herbal prescription recommendation, realizing the integrated processing from patients’ clinical text information to syndrome type identification, disease diagnosis, and personalized prescription generation. The system adopts a pre-trained language model (BERT) and a multi-task learning architecture to synchronously complete three core tasks: syndrome differentiation, disease identification, and prescription recommendation. Supported by 1500 real desensitized medical records, the experiment divides the dataset into training, validation, and test set at a ratio of 8:2:5, and evaluates model performance through multi-dimensional indicators. The results show that the system achieves a comprehensive accuracy of 0.52 in the syndrome differentiation and disease identification task and a comprehensive score of 0.48 in the prescription recommendation task, with overall performance superior to traditional single-task models and the baseline model. This study provides a referable technical solution for intelligent TCM diagnosis and treatment, which helps improve the efficiency and standardization of primary-level diagnosis and treatment and promotes the modernization of TCM.