基于动态多维时序数据的高阶收敛LM神经网络构建名中医辨证论治模型研究—以特发性肺纤维化为例

特发性肺纤维化（IPF）属病因不明的呼吸系统难治性疾病，中医特别是名(老)中医诊疗对改善患者临床症状有明显效果。辨证论治是中医学的精髓，证候是其中的关键要素，如何将名医的宝贵诊疗经验客观化？课题组前期对名中医治疗IPF文献医案（116例）和首届全国名中医张之文治疗该病的回顾性病例（79例）进行数据挖掘研究（仅选取首诊次），发现名医诊疗该病在辨证和论治方面存在一定的共性规律。慢性难治疾病的真实临床诊疗呈现多诊次（不同时序）证候发展变化的特点，医家的“论治”也会随之变化，不同时间序列的辨证和论治是否有变化规律可循？结合前期研究基础，本项目拟运用多维关联规则分析影响IPF证候分类的核心症状和辨治不同证候的核心方药，通过收敛速度更快、拟合度更优和预测精度更高的高阶收敛LM神经网络，基于多维时序数据构建“症-证-治-方”模型，探索名中医辨证论治的动态变化规律，提升中医药防治重大疑难疾病的能力。

Idiopathic Pulmonary Fibrosis (IPF) is a refractory disease of respiratory system with unknown etiology. The diagnosis and treatment of TCM physicians, especially famous(old) TCM physicians have noticeable effect in improving patients' clinical symptoms. Syndrome differentiation and treatment are the essences of TCM and syndromes are the key elements. How do we objectify the valuable experience in diagnosis and treatment of famous TCM physicians? Through data mining, our research team selectively studied the first-diagnosis cases of 116 cases in medical literature of the treatment of IPF by famous TCM physicians and 79 retrospective cases treated by Dr. Zhang Zhiwen who is one of the first national famous TCM physicians. We found that there are common rules in syndrome differentiation and treatment of famous TCM physicians. The real clinical diagnosis and treatment of chronic refractory disease have multi-diagnosis and different time-series development of syndromes and the 'diagnosis and treatment' change accordingly. Are there rules to follow for the syndrome differentiation and treatment with different time series? On the basis of previous research, this project intends to discover the core symptoms affecting the syndrome classification of IPF and core prescription for differentiating and treating different syndromes, construct the model of syndrome differentiation and treatment of ‘symptom-syndrome-treatment-prescription’ based on the dynamic and multi-dimensional time-series data through L-M neural network of higher order convergence with higher speed of convergence, better degree of fitting and higher prediction accuracy. Thus, this project intends to explore the dynamic change rules of TCM syndrome differentiation and treatment and improve the dynamic change rules of TCM syndrome differentiation and treatment and improves the abilities for prevention and treatment of major and difficult diseases with Chinese medicine.