数据挖掘技术在肿瘤疾病诊疗中的应用研究

发布时间：2018-03-26 01:19

本文选题：数据预处理　切入点：算法对比　出处：《青岛科技大学》2017年硕士论文

【摘要】：本文主要进行肿瘤疾病危险因素、发病特征、治疗过程和治疗方式的挖掘分析,并进行主流挖掘算法的性能的对比,从而得到最合适的疾病诊断与治疗的辅助挖掘算法。通过对医院信息化系统中存储的肿瘤数据进行整理,做数据预处理和筛选主要关键词,最终得到适合挖掘的数据形式,然后采用数据挖掘算法对数据进行挖掘操作,得到潜在价值、规律和算法性能。通过实验最终得到以下结果:(1)对于主流挖掘算法的原理和特性做了比较。根据在医学领域的使用现状,分析了几种主流挖掘算法在医学领域的适用范围和协同使用的可行性。(2)对于挖掘结果做了对比分析,根据算法性能进行挖掘算法的选择,最终选择最优的挖掘算法进行肿瘤疾病的数据分析。(3)通过对于肿瘤致病因素、治疗方法(手术和用药)和疾病间的关联规则进行挖掘分析,明确关联强度,可以为基础实验和临床研究提供提示。(4)通过对肿瘤疾病危险因素、发病特征、治疗过程和治疗手段的挖掘分析来帮助医生进行肿瘤疾病的辅助诊断和治疗。最终,本文通过对肿瘤类疾病领域数据挖掘技术使用的现状进行综合分析,明确了数据挖掘技术的概念、原理和在医学领域中的运用趋势,并且总结了几种主流数据挖掘方法的适用范围。通过对数据挖掘方法的研究发现,可根据挖掘算法的自身特点,来选用最合适的挖掘方法。在通过数据的处理过程中,研究了如何科学的进行数据的属性提取,降维、降噪处理,以及数据的类型转换和缺失数据补充等操作,对于数据的预处理方式有一定探索和指导意义。在通过数据的挖掘实验中得到了肿瘤疾病的一些关联规则,对于肿瘤疾病的认识、预防和治疗具有一定的指导意义。最后,通过对比主流的挖掘算法的使用结果,分析了各个算法的自身特点和适用特点,并且探索了统一医疗信息系统的创建方式。
[Abstract]:In this paper, the risk factors of tumor disease, the characteristics of the disease, the course of treatment and the treatment method are analyzed, and the performance of the mainstream mining algorithm is compared. By sorting out the tumor data stored in the hospital information system, making data preprocessing and screening the main keywords, finally obtaining the data form suitable for mining. Then the data mining algorithm is used to mine the data, and the potential value is obtained. The principle and characteristics of the mainstream mining algorithms are compared by the following results: 1: 1. According to the current situation of application in the field of medicine, this paper makes a comparison of the principles and characteristics of the mainstream mining algorithms. This paper analyzes the scope of application of several mainstream mining algorithms in medical field and the feasibility of collaborative use. (2) the mining results are compared and analyzed, and the selection of mining algorithms is carried out according to the performance of the algorithm. Finally, the optimal mining algorithm is selected to analyze the data of tumor diseases. (3) by mining and analyzing the association rules between tumor pathogenic factors, treatment methods (surgery and medication) and diseases, the association intensity is clear. It can provide hints for basic experiments and clinical studies to help doctors with the auxiliary diagnosis and treatment of tumor diseases by digging and analyzing the risk factors, characteristics, treatment process and treatment methods of tumor diseases. Based on the comprehensive analysis of the current situation of the application of data mining technology in the field of tumor diseases, this paper clarifies the concept, principle and application trend of data mining technology in the field of medicine. Through the research of the data mining method, we find that the most suitable mining method can be selected according to the characteristics of the mining algorithm. In this paper, we study how to scientifically carry out data attribute extraction, dimensionality reduction, noise reduction, data type conversion and missing data supplement, etc. In the data mining experiment, some association rules of tumor diseases are obtained, which have certain guiding significance for the understanding, prevention and treatment of tumor diseases. By comparing the results of the mainstream mining algorithms, this paper analyzes the characteristics and applicable characteristics of each algorithm, and explores the way to create a unified medical information system.
【学位授予单位】：青岛科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R730;TP311.13

【参考文献】