Research And Application Of Spatio-Temporal Correlation Analysis Method

With the development of information technology and communication technology, digital storage technology and high-speed data acquisition technology, a lot of data with spatial and temporal properties have gradually accumulated in transportation, electricity, logistics, environmental monitoring, industrial production. The explosive data, which could be acquired at anytime and anywhere, bring us a lot of information; however, the extrusive problem"massive data makes a hard knowledge discovery process"become more and more serious. Recently, more and more researchers realize that researching spatial objects over time variation, to discovery spatio-temporal association rules ,and using those rules to analyze the trend of spatio-temporal variation and predict the future of spatio-temporal state have important application value for planning and construction, command scheduling, emergency management and information services and other.The purpose of this study is to find spatio-temporal association rules by using spatio-temporal association rule algorithm. Using this rules to analyze the trend of traffic congestion. With the knowledge discovered from traffic congestion, we can provide more useful information for route navigation, trend information and traffic control.The aim of spatio-temporal correlation analysis is to study spatial objects over time variation, in order to reflect the data of association in time and space. Spatio-temporal association rules mining is one of the main methods of spatio-temporal correlation analysis, which has been conducted research or application by many scholars. This thesis describes the current research status towards spatio-temporal association rules mining, and proposes a new spatio-temporal association rules mining method by analyzing the existing methods taking into account the lack of spaitial and temporal constrains. We classify the time duration of spatio-temporal data and consider the spatial relationship firstly and generates the transaction table, then performs join operation on spatial-related item sets. During the executing of the algorithm, association rules related threshold have been analyzed, so that the mining results to better meet the needs of users. The analysis and experiments illuminate that the algorithm is well performed. Taking into account the spatial and temporal constraints simultaneously can filter irrelevant data early and improve the efficiency of discovering spatio-temporal association rule. One the basis of the theory, this thesis finally design and implement a system to analyze the trend of traffic congestion based on spatio-temporal correlation analysis method. This system can effectively analysis spatio-temporal correlation and visualize the results.