【摘要】：現(xiàn)如今，空間數(shù)據(jù)和信息技術(shù)的不斷發(fā)展，地理信息系統(tǒng)受到越來越多的關(guān)注。由于空間數(shù)據(jù)探測(cè)技術(shù)的發(fā)展，人們掌握了多種的獲取數(shù)據(jù)的方法，因此海量的空間位置相關(guān)的數(shù)據(jù)被人們所逐步積累。因此，人們需要一個(gè)執(zhí)行力強(qiáng)的數(shù)據(jù)分析工具來實(shí)現(xiàn)從空間數(shù)據(jù)庫中獲取知識(shí)和信息。而數(shù)據(jù)挖掘技術(shù)作為一種新型的數(shù)據(jù)分析技術(shù)可以發(fā)現(xiàn)數(shù)據(jù)庫數(shù)據(jù)的潛在的價(jià)值信息知識(shí)，在這種情況下，空間數(shù)據(jù)挖掘技術(shù)應(yīng)運(yùn)而生。 所謂數(shù)據(jù)挖掘就是從表面看毫無規(guī)律的數(shù)據(jù)中提取出有價(jià)值的知識(shí)。現(xiàn)如今的數(shù)據(jù)量很大，人們希望從這些數(shù)據(jù)中獲得到知識(shí)，所以這中對(duì)數(shù)據(jù)的處理技術(shù)受到廣泛關(guān)注。數(shù)據(jù)挖掘涉及到的技術(shù)很多，其中分類預(yù)測(cè)是一種常見的技術(shù)。數(shù)據(jù)挖掘技術(shù)中涵蓋多種算法，決策樹算法是一種優(yōu)勢(shì)很明顯的算法。它通過歸納算法對(duì)數(shù)據(jù)進(jìn)行分類，計(jì)算任務(wù)不大，規(guī)則也很明顯，得到廣泛應(yīng)用。 本課題主要完成的工作量以及研究內(nèi)容主要包括以下幾個(gè)方面： （1）詳述本課題的研究背景及研究意義，還介紹了決策樹算法在國內(nèi)外的研究現(xiàn)狀以及地理信息系統(tǒng)的發(fā)展和研究現(xiàn)狀。 （2）概述數(shù)據(jù)挖掘、數(shù)據(jù)挖掘系統(tǒng)、空間數(shù)據(jù)挖掘的概念，并且著重介紹了分類預(yù)測(cè)的概念。同時(shí)，對(duì)當(dāng)前存在的多種分類預(yù)測(cè)方法進(jìn)行了簡(jiǎn)單介紹。 （3）系統(tǒng)的對(duì)決策樹算法進(jìn)行了詳細(xì)地討論。闡述了決策樹算法的概述、構(gòu)造、簡(jiǎn)化以及決策樹的性能評(píng)價(jià)、算法實(shí)現(xiàn)。同時(shí)討論了決策樹算法中的幾種常用的典型的方法和比較，以及部分算法的程序?qū)崿F(xiàn)。還對(duì)決策樹算法的常見問題進(jìn)行了討論。 （4）對(duì)決策樹算法在GIS中的應(yīng)用進(jìn)行了研究。本文采用了土地宜耕性的例子，根據(jù)空間數(shù)據(jù)應(yīng)用決策樹算法中的兩種方法進(jìn)行數(shù)據(jù)挖掘，并建立空間數(shù)據(jù)的決策樹，同時(shí)評(píng)價(jià)兩種方法，以及應(yīng)用決策樹進(jìn)行預(yù)測(cè)，并構(gòu)建實(shí)例。 （5）本文最后對(duì)決策樹算法在GIS中的研究進(jìn)行了總結(jié)與展望。對(duì)自己在這個(gè)課題中完成的工作做出了相關(guān)的總結(jié)，，也對(duì)自己的不足之處提出了改進(jìn)希望。同時(shí)也對(duì)地理信息的發(fā)展也做出來展望。
[Abstract]:Nowadays, with the development of spatial data and information technology, GIS attracts more and more attention. With the development of spatial data detection technology, people have mastered a variety of methods to obtain data, so massive spatial position related data is gradually accumulated by people. Therefore, people need a powerful data analysis tool to obtain knowledge and information from spatial database. As a new kind of data analysis technology, data mining technology can discover the potential value information knowledge of database data. In this case, spatial data mining technology emerges as the times require. Data mining is to extract valuable knowledge from seemingly irregular data. Nowadays, the amount of data is very large, people want to get knowledge from these data, so the technology of data processing has been paid more and more attention. There are many techniques involved in data mining, among which classification and prediction is a common technique. Data mining technology covers many algorithms, decision tree algorithm is a very obvious advantage of the algorithm. It classifies the data by inductive algorithm, and the calculation task is not large and the rules are obvious, so it is widely used. The workload and research contents of this subject mainly include the following aspects: (1) the research background and significance of this subject are described in detail. The research status of decision tree algorithm at home and abroad and the development and research status of GIS are also introduced. (2) the concepts of data mining, data mining system and spatial data mining are summarized, and the concepts of classification and prediction are introduced emphatically. At the same time, several existing classification and prediction methods are briefly introduced. (3) the decision tree algorithm is discussed in detail. The summary, construction, simplification, performance evaluation and implementation of decision tree algorithm are described in this paper. At the same time, several typical methods and comparisons of decision tree algorithms are discussed, and some of the algorithms are implemented. The common problems of decision tree algorithm are also discussed. (4) the application of decision tree algorithm in GIS is studied. In this paper, an example of land suitability for ploughing is used to mine data according to two methods of spatial data application decision tree algorithm, and to establish decision tree of spatial data. At the same time, two methods are evaluated, and decision tree is applied to forecast. And build an example. (5) in the end, the research of decision tree algorithm in GIS is summarized and prospected. This paper makes a summary of the work done in this subject, and puts forward some suggestions for improvement. At the same time, the development of geographic information is also expected.
【學(xué)位授予單位】：中國地質(zhì)大學(xué)（北京）
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2013
【分類號(hào)】：P208;TP311.13

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 王靜紅,李筆;基于決策樹的一種改進(jìn)算法[J];電訊技術(shù);2004年05期

2 何凡能,葛全勝,鄭景云;中國清代城鎮(zhèn)用地面積估算及其比較[J];地理學(xué)報(bào);2002年06期

3 何凡能;葛全勝;戴君虎;林珊珊;;近300年來中國森林的變遷[J];地理學(xué)報(bào);2007年01期

4 曲開社,成文麗,王俊紅;ID3算法的一種改進(jìn)算法[J];計(jì)算機(jī)工程與應(yīng)用;2003年25期

5 王曉國,黃韶坤,朱煒,李啟炎;應(yīng)用C4.5算法構(gòu)造客戶分類決策樹的方法[J];計(jì)算機(jī)工程;2003年14期

6 欒麗華,吉根林;決策樹分類技術(shù)研究[J];計(jì)算機(jī)工程;2004年09期

7 李志聰;鐘紹春;;自適應(yīng)學(xué)習(xí)系統(tǒng)中樸素貝葉斯分類的應(yīng)用研究[J];計(jì)算機(jī)工程與科學(xué);2009年09期

8 萬文;;礦井突水水源分析的Bayes判別分析模型及其應(yīng)用[J];礦業(yè)工程研究;2009年03期

9 劉小虎,李生;決策樹的優(yōu)化算法[J];軟件學(xué)報(bào);1998年10期

10 路紅梅;;基于決策樹的經(jīng)典算法綜述[J];宿州學(xué)院學(xué)報(bào);2007年02期

相關(guān)碩士學(xué)位論文 前4條

1 楊緒兵;線性判別分析及其推廣性研究[D];南京航空航天大學(xué);2003年

2 周旋;基于貝葉斯網(wǎng)絡(luò)的作戰(zhàn)效能評(píng)估與分析方法研究[D];國防科學(xué)技術(shù)大學(xué);2008年

3 曾志中;基于貝葉斯算法的垃圾郵件過濾系統(tǒng)的分析與實(shí)現(xiàn)[D];北京郵電大學(xué);2009年

4 李永;分類算法研究[D];吉林大學(xué);2009年

本文編號(hào)：2369258