本文選題：數(shù)據(jù)預(yù)處理　切入點(diǎn)：算法對(duì)比　出處：《青島科技大學(xué)》2017年碩士論文

【摘要】：本文主要進(jìn)行腫瘤疾病危險(xiǎn)因素、發(fā)病特征、治療過程和治療方式的挖掘分析,并進(jìn)行主流挖掘算法的性能的對(duì)比,從而得到最合適的疾病診斷與治療的輔助挖掘算法。通過對(duì)醫(yī)院信息化系統(tǒng)中存儲(chǔ)的腫瘤數(shù)據(jù)進(jìn)行整理,做數(shù)據(jù)預(yù)處理和篩選主要關(guān)鍵詞,最終得到適合挖掘的數(shù)據(jù)形式,然后采用數(shù)據(jù)挖掘算法對(duì)數(shù)據(jù)進(jìn)行挖掘操作,得到潛在價(jià)值、規(guī)律和算法性能。通過實(shí)驗(yàn)最終得到以下結(jié)果:(1)對(duì)于主流挖掘算法的原理和特性做了比較。根據(jù)在醫(yī)學(xué)領(lǐng)域的使用現(xiàn)狀,分析了幾種主流挖掘算法在醫(yī)學(xué)領(lǐng)域的適用范圍和協(xié)同使用的可行性。(2)對(duì)于挖掘結(jié)果做了對(duì)比分析,根據(jù)算法性能進(jìn)行挖掘算法的選擇,最終選擇最優(yōu)的挖掘算法進(jìn)行腫瘤疾病的數(shù)據(jù)分析。(3)通過對(duì)于腫瘤致病因素、治療方法(手術(shù)和用藥)和疾病間的關(guān)聯(lián)規(guī)則進(jìn)行挖掘分析,明確關(guān)聯(lián)強(qiáng)度,可以為基礎(chǔ)實(shí)驗(yàn)和臨床研究提供提示。(4)通過對(duì)腫瘤疾病危險(xiǎn)因素、發(fā)病特征、治療過程和治療手段的挖掘分析來幫助醫(yī)生進(jìn)行腫瘤疾病的輔助診斷和治療。最終,本文通過對(duì)腫瘤類疾病領(lǐng)域數(shù)據(jù)挖掘技術(shù)使用的現(xiàn)狀進(jìn)行綜合分析,明確了數(shù)據(jù)挖掘技術(shù)的概念、原理和在醫(yī)學(xué)領(lǐng)域中的運(yùn)用趨勢(shì),并且總結(jié)了幾種主流數(shù)據(jù)挖掘方法的適用范圍。通過對(duì)數(shù)據(jù)挖掘方法的研究發(fā)現(xiàn),可根據(jù)挖掘算法的自身特點(diǎn),來選用最合適的挖掘方法。在通過數(shù)據(jù)的處理過程中,研究了如何科學(xué)的進(jìn)行數(shù)據(jù)的屬性提取,降維、降噪處理,以及數(shù)據(jù)的類型轉(zhuǎn)換和缺失數(shù)據(jù)補(bǔ)充等操作,對(duì)于數(shù)據(jù)的預(yù)處理方式有一定探索和指導(dǎo)意義。在通過數(shù)據(jù)的挖掘?qū)嶒?yàn)中得到了腫瘤疾病的一些關(guān)聯(lián)規(guī)則,對(duì)于腫瘤疾病的認(rèn)識(shí)、預(yù)防和治療具有一定的指導(dǎo)意義。最后,通過對(duì)比主流的挖掘算法的使用結(jié)果,分析了各個(gè)算法的自身特點(diǎn)和適用特點(diǎn),并且探索了統(tǒng)一醫(yī)療信息系統(tǒng)的創(chuàng)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.
【學(xué)位授予單位】：青島科技大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：R730;TP311.13

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 童剛;王魯興;董一鴻;;醫(yī)院資產(chǎn)管理系統(tǒng)的優(yōu)化[J];中國(guó)數(shù)字醫(yī)學(xué);2016年07期

2 牟冬梅;馮超;王萍;;數(shù)據(jù)挖掘方法在醫(yī)學(xué)領(lǐng)域的應(yīng)用及SWOT分析[J];醫(yī)學(xué)信息學(xué)雜志;2015年01期

3 杜巍;趙春榮;黃偉建;;改進(jìn)的k-means聚類算法在客戶細(xì)分中的應(yīng)用研究[J];河北經(jīng)貿(mào)大學(xué)學(xué)報(bào);2014年01期

4 魏小玲;譚善娟;何其棟;王威;吳擁軍;王靜;吳逸明;;決策樹聯(lián)合生物標(biāo)志在肺癌輔助診斷中應(yīng)用[J];中國(guó)公共衛(wèi)生;2013年10期

5 蔡佳慧;張濤;宗文紅;;醫(yī)療大數(shù)據(jù)面臨的挑戰(zhàn)及思考[J];中國(guó)衛(wèi)生信息管理雜志;2013年04期

6 李翔;朱全銀;;Adaboost算法改進(jìn)BP神經(jīng)網(wǎng)絡(luò)預(yù)測(cè)研究[J];計(jì)算機(jī)工程與科學(xué);2013年08期

7 楊佳琦;陳露菲;陳淑紅;邵冰;曹博;李航;焦U，

本文編號(hào)：1665700