决策树算法在GIS中的应用研究

发布时间：2018-12-09 10:50

【摘要】：现如今，空间数据和信息技术的不断发展，地理信息系统受到越来越多的关注。由于空间数据探测技术的发展，人们掌握了多种的获取数据的方法，因此海量的空间位置相关的数据被人们所逐步积累。因此，人们需要一个执行力强的数据分析工具来实现从空间数据库中获取知识和信息。而数据挖掘技术作为一种新型的数据分析技术可以发现数据库数据的潜在的价值信息知识，在这种情况下，空间数据挖掘技术应运而生。所谓数据挖掘就是从表面看毫无规律的数据中提取出有价值的知识。现如今的数据量很大，人们希望从这些数据中获得到知识，所以这中对数据的处理技术受到广泛关注。数据挖掘涉及到的技术很多，其中分类预测是一种常见的技术。数据挖掘技术中涵盖多种算法，决策树算法是一种优势很明显的算法。它通过归纳算法对数据进行分类，计算任务不大，规则也很明显，得到广泛应用。本课题主要完成的工作量以及研究内容主要包括以下几个方面：（1）详述本课题的研究背景及研究意义，还介绍了决策树算法在国内外的研究现状以及地理信息系统的发展和研究现状。（2）概述数据挖掘、数据挖掘系统、空间数据挖掘的概念，并且着重介绍了分类预测的概念。同时，对当前存在的多种分类预测方法进行了简单介绍。（3）系统的对决策树算法进行了详细地讨论。阐述了决策树算法的概述、构造、简化以及决策树的性能评价、算法实现。同时讨论了决策树算法中的几种常用的典型的方法和比较，以及部分算法的程序实现。还对决策树算法的常见问题进行了讨论。（4）对决策树算法在GIS中的应用进行了研究。本文采用了土地宜耕性的例子，根据空间数据应用决策树算法中的两种方法进行数据挖掘，并建立空间数据的决策树，同时评价两种方法，以及应用决策树进行预测，并构建实例。（5）本文最后对决策树算法在GIS中的研究进行了总结与展望。对自己在这个课题中完成的工作做出了相关的总结，，也对自己的不足之处提出了改进希望。同时也对地理信息的发展也做出来展望。
[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.
【学位授予单位】：中国地质大学（北京）
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：P208;TP311.13

【参考文献】