数据挖掘技术在肺结核疾病诊断中的应用研究

发布时间：2018-05-07 22:39

本文选题：肺结核疾病 + 粗糙集　；参考：《吉林农业大学》2017年硕士论文

【摘要】：结核病多年来一直严重威胁着各国人民的身体健康。我国是全球22个结核病高负担国家之一。信息技术的不断进步,不断发展,各大医院中中也都使用医疗管理系统进行日常工作,使得其在医疗设备和仪器的数字化后,信息量也在不断的增长扩大。如何通过高效、智能的计算机算法对海量肺结核疾病诊疗数据进行数据挖掘,并进行智能诊断,是本文研究内容。通过收集整理北京市昌平区结核病防治所、北京市结核病控制研究所的8210例肺结核病人档案,应用数据库技术,构建基于SQL Server 2010的疾病电子档案,运用数据挖掘方法中的粗糙集和决策树方法,建立肺结核疾病智能诊断模型。考虑到医学大数据及数据共享,本文搭建了基于hadoop的大数据平台,构建肺结核疾病诊断智能诊断云系统,以满足医学大数据的需求。本文的研究内容及主要工作包含以下4点:1、本文对数据集进行预处理后,优化了数据集,以应对医学数据中的存在噪声、不完全的问题。本文选择数据挖掘中的聚类方法对肺结核疾病数据进行预处理;保证数据的一致性和不确定性,并将其转化为适合挖掘的形式。该项操作,能够很好地进行后期高质量的数据挖掘。2、肺结核疾病因个体不同所表现出的症状也有差异,数据属性众多,导致挖掘效率不高。针对这一问题,本文运用粗糙集理论进行属性约简,并提出适用于SQL语言条件信息熵计算方法,完成属性约简。利用基于数据库的粗糙集属性约简方法,去除了肺结核疾病数据中的冗余属性。3、在模型构建中,针对单一挖掘方法挖掘效率不高的问下,本研究应用粗糙集和决策树相结合的方法完成模型构建。对比使用单一的决策树方法和粗糙集方法,本文提出的粗糙集约简与决策树规则提取相融合的优化算法,具有良好的适用性及优势,在保证诊断的客观性的前提下,有效的提高了肺结核疾病诊断的准确性。4、为满足不断增长的医学大数据的需求,本文在构建肺结核疾病诊断模型的基础上,应用hadoop技术搭建大数据云平台,并建立肺结核疾病智能诊断系统,将云计算应用到医疗系统当中,有效地处理医疗数据,为肺结核疾病诊断提供决策支持。
[Abstract]:Tuberculosis has been a serious threat to the health of people all over the world for many years. China is one of the 22 countries with high TB burden. With the continuous progress and development of information technology, medical management system is also used in every hospital for daily work, which makes the amount of information increase and expand after the digitization of medical equipment and instruments. How to mine the data of massive tuberculosis diagnosis and treatment data and make intelligent diagnosis by efficient and intelligent computer algorithm is the research content of this paper. Through collecting and sorting out the records of 8210 cases of tuberculosis patients from the Institute of Tuberculosis Control and Prevention in Changping District, Beijing, and applying the database technology, the electronic archives of diseases based on SQL Server 2010 were constructed. The intelligent diagnosis model of pulmonary tuberculosis is established by using rough set and decision tree method in data mining. Considering the medical big data and the data sharing, this paper builds an intelligent diagnostic cloud system based on hadoop for diagnosis of pulmonary tuberculosis, in order to meet the needs of medical big data. The research contents and main work of this paper include the following 4 points: 1. After preprocessing the data set, we optimize the data set to deal with the problem of noise and incomplete in medical data. In this paper, we choose the clustering method in data mining to preprocess the data of tuberculosis disease, ensure the consistency and uncertainty of the data, and transform it into a form suitable for mining. This operation can well carry out high quality data mining in the later period. The symptoms of pulmonary tuberculosis disease are different because of the different individuals, and the data attributes are numerous, which leads to the low efficiency of mining. In order to solve this problem, this paper applies rough set theory to attribute reduction, and proposes a method for computing conditional information entropy in SQL language to complete attribute reduction. Using the attribute reduction method of rough set based on database, the redundant attribute. 3 in the tuberculosis disease data is removed. In the model construction, the mining efficiency of the single mining method is not high. In this study, rough set and decision tree are used to build the model. Compared with the single decision tree method and rough set method, this paper proposes an optimization algorithm which combines rough set reduction and decision tree rule extraction, which has good applicability and advantages, and ensures the objectivity of diagnosis. The accuracy of pulmonary tuberculosis diagnosis is improved effectively. In order to meet the increasing demand of medical big data, this paper builds a diagnosis model of pulmonary tuberculosis, and applies hadoop technology to build the big data cloud platform. The intelligent diagnosis system of tuberculosis disease is established and cloud computing is applied to the medical system to deal with the medical data effectively and to provide decision support for the diagnosis of tuberculosis disease.
【学位授予单位】：吉林农业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R521;TP311.13

【相似文献】