Research Of Spatio-temporal Data Mining Algorithm Based On FP-tree

Spatio-temporal Data Mining is an extension of Spatial Data Mining. It is mainly aimed at the analysis and processing of spatio-temporal data. The structure of spatio-temporal data includes three aspects: location information, time information and attribute information. Spatio-temporal Data Mining is not only to analyze the attribute information, but also need to analyze the spatial relationship and the variational rules with time, and then excavate the inner link which hidden behind the spatio-temporal data between time and space in the real world, so as to provide the help of the space decision in objective world.The paper is mainly to research the spatio-temporal data mining algorithms. Firstly, the basic concepts of Data mining, Spatial Data mining, Spatio-temporal Data Mining and the difference between the main algorithms are discussed. Secondly, focus on the analysis and exploration of traditional association rules mining algorithms and to find their problems. Finally, to overcome the defects of classic association rules mining algorithms, a spatio-temporal association rules mining algorithm based on FP-tree(STFP-tree)is proposed.Experiments demonstrate that STFP-tree algorithm can effectively analyze and dig out spatio-temporal data, and realize the transformation from the traditional association rules algorithm to the spatial-temporal correlation algorithm. Its application in the trend analysis and prediction of the hotel accommodation prices around schools achieved the desired effects and could provide the decision basis for the hotel location choice and the hotel price adjustment.