基于频繁项集挖掘的LUCC轨迹分析

Frequent Itemset Mining Based LUCC Trajectory Analysis

针对多时相、多类别情况下土地利用/覆被变化轨迹提取难与分析难的问题,提出一种基于频繁项集挖掘的土地利用/覆被变化轨迹分析方法。该方法首先使用FI-Apriori算法对土地利用/覆被数据进行频繁项集挖掘,然后将挖掘结果用于土地利用/覆被变化轨迹分析。采用美国地质勘探局土地覆被数据的实验表明该方法可行且高效,其结果信息量大且能以易判读的形式进行表述,有利于实现变化规律的分析和总结。

A novel land use and land cover change trajectory extract and analysis method based on frequent itemsets mining is proposed in this paper. This method use FI- Apriori,a spatial data mining algorithm designed by us,to mine spatial frequent itemsets from multi- temporal and multi- type land use and land cover data efficiently; with these frequent itemsets,area statistics,transfer patterns and change trajectories of land use and land cover types can be easily fetched and analyzed. Experiment using data fetched by USGS shows this method is valid and efficient in land use and land cover trajectory analysis; results of the analysis contain much information,and can be expressed in highly interpretable forms.