基于决策树的耕地转建设用地分析与预测——以北京房山区为例

Analysis and Forecast of Land Change from Cultivated Land to Construction Land Based on Decision Trees: A Case Study of Fangshan District,Beijing

利用C5.0决策树算法对房山区耕地转变为建设用地情况进行了分析和预测。首先,利用房山区的土地利用数据、自然和社会经济因子数据进行数据挖掘,产生了房山区耕地转建设用地的空间关联规则,反映了房山区耕地转为建设用地的主要驱动因素和变化规律,表明社会经济因子在耕地转建设用地中起主要作用。其次,利用转换规则的置信度信息,对房山区耕地转为建设用地的转换概率进行了预测,生成了概率分析图,分析和预测了房山区可能发生耕地转建设用地的热点地段。结果表明,该文提出的方法能够有效分析和预测房山区耕地转变为建设用地的情况,对于其他地区的土地利用研究也具有一定的借鉴意义。

The paper analyzes and forecasts the condition of cultivated land changing to construction land in Fangshan District（in Beijing） using C5.0 decision tree algorithm. Firstly, the paper produces the spatial associate rules from cultivated land to construction land in Fangshan, using data mining methods with the data of Fangshan land use, natural factors and social-economic factors. The changing rules can reflect the main driving factors and changing disciplines on changing from cultivated land to construction land,and they prove that social-economic factors can play a dominating part in transforming cultivated land to construction land. Moreover, the authors use the information of confidence level of changing rules, forecast and analyze the changing probability of changing from cultivated land to construction land in Fangshan from the prospective of probability analysis. The analysis produces the map of probability analysis, and the map shows the hotspot areas which may have land changing from cultivated land to construction land. Through the test and verifying of experiment results, it is proved that the methods of this paper can effectively analyze and forecast the condition of land changing from cultivated land to construction land. Moreover, the paper also has some value of reference for research on other areas.