摘要对于帕金森病，中医对患者的证型进行诊断的基本手段有望闻问切等，同时联系中医量表。医生的经验和中医量表的制定直接影响了医生对患者的诊断。因此，在对帕金森病的诊断过程中，如何尽量减少或者避免证型误判以及如何提高中医量表的规范化和客观化成为主要难题。82525

在中医诊断帕金森病的过程中，可以将症状属性作为特征，可能的证型作为标签集，那么就可以将多标签学习引入到中医诊断帕金森病的领域中，同时用于规范和客观中医量表以及提高诊断率。将多标签学习与传统中医结合，推动中医量表的规范化和客观化，促进中医理论体系的完备，自然也是一个新颖且有价值的研究方向。

本文尝试采用多标签分类技术来构建中医治疗帕金森病的模型，并将其应用到中医诊断中，从而可以帮助中医在治疗帕金森病领域的发展。本文的主要思路是讨论多标签学习的研究现状，挑选算法帕金森数据集上进行实现并进行预测性能的比较，最终挑选出较适合帕金森数据集的算法；然后制作中医诊断帕金森病辅助系统，将之前挑选的算法作为系统的数据分析部分。

本文的工作如下：

(1)  主要讨论了中医及其诊断帕金森病的现状和多标签分类用于中医领域的可行性。

(2)  详细介绍了当前多标签学习领域的研究现状。

(3)  探索和比较不同多标签算法在帕金森数据集上的执行效果和特点。

(4)  制作以多标签分类算法为核心的中医诊断帕金森病的预测系统。

(5)  总结已有工作，展望未来多标签分类的可研究前景。

文章研究了多标签分类算法，制作了预测系统，可以看作是多标签分类在中医诊断帕金森病的一次尝试，希望能推动相关领域的研究。

毕业论文关键词  多标签学习   标签相关性   帕金森   中医   应用

毕业设计说明书外文摘要

Title   Study on Multi-label Learning and its application in Traditional Chinese Medicine diagnosing Parkinson disease

Abstract For Parkinson disease, Traditional Chinese Medical(TCM) uses four basic methods of Chinese diagnosis(Observation,Olfaction,Inquiry and Palpation) combined with TCM scale to diagnose Parkinson。The doctors' experience and the formulation of TCM scale affect the diagnosis of patients directly。Therefore,how to reduce or avoid the misdiagnosis and how to improve the standardization and objectification of TCM scale are major problems in the diagnosis of Parkinson disease in TCM。

Regarding symptoms and syndromes as features and labels,we can introduce Multi-label Learning(MLL) to the diagnosis of Parkinson disease in TCM ,which also promotes the standardization and objectification of TCM scale and improves the accuracy of the diagnosis。The combination of MLL and TCM promotes the standardization and objectification of TCM scale ,and promotes the perfection of the theoretical system of TCM,which is a novel and valuable research direction。

This paper attempts to use MLL to build a predictive medical system which can be used to diagnose Parkinson disease in TCM and can promote the development of TCM。The paper aims to provide an introduction to MLL,compare the performance of some algorithms of MLL running on Parkinson data set and select the best prediction algorithm as the data analysis part in auxiliary system of TCM diagnosis of Parkinson。

This paper mainly focuses on the follow aspects:

(1) Discusses the current research status of TCM ,the diagnosis of Parkinson disease in TCM and the feasibility of combination of MLL and TCM。

(2) Introduces the current research status of MLL in detail。