Main Article Content
With the development of remote sensing technology and the commercialization of aerospace industry, the requirements of remote sensing observation will grow rapidly in the future, and the shortage of remote sensing resources and the conflict between observation tasks will become increasingly prominent. For the lack of mature data mining means for remote sensing user demand analysis and demand pre-processing, this paper proposes a
novel association analysis method based on directed term graph-based double-layer FP-tree to extract users' remote sensing demands and observation target characteristics. Firstly, by analyzing the dependencies between the demand items, the directed term graph about user, target, payload and observation date is designed. Secondly, a double-layer FP-tree based on the directed term graph is presented. Such a tree can effectively express the logical
relationship between demand data and realize the data compression of common input parts, which features small data amount and fast operation. And then, an association rule analysis method based on FP-growth algorithm for remote sensing user demand mining is proposed. In the process of association analysis, the double-layer FP-tree is decomposed into four sub-FP-trees, and for each sub-FP-tree, FP-growth algorithm is adopted to generate frequent itemset. This frequent itemset generation process has some superior properties, i.e., the data set is scanned only twice, the search path is short, search range is small, and there is no need to generate candidate itemset, and no repeated frequent itemset is generated. Finally, a transaction set of remote sensing demands is analyzed to verify the effectiveness of the proposed method. Some conclusions are drawn in the end of this paper.