| GML is based on XML, the spatial information coding standard OpenGIS Consortium (OGC), get the Oracle, and put forward the famous company ESRI MapInfo, support. Text is simple, intuitive and easy to understand and edit. GML exploits text these advantages, to express describe geographic information knowledge. Currently, GML technology research made a lot of achievements, such as: GML query, GML storage, GML analytical, GML visualization, etc. GML tense model describing geometry, topology, tense reference system and geographical data tense features components. Basic space-time model is committed to providing elements layer and attribute layer, and to support the timestamp tracking targets. State and event is of space time data model two basic model.Data mining is from a lot of, incomplete, noisy, fuzzy and random practical application in the data, the extract implicit in which people prior don't know, but also potentially useful information and knowledge of the process.Spatio-temporal data has time, space, and the space characteristics. The spatio-temporal data exist space and time scale; Existing space and time the relations; Existing spatio-temporal correlation and heterogeneous space.Outliers mining in data mining areas is an important part of the research in many fields, research outliers more useful than research clustering, more important. On many occasions, found outliers has a very important significance. Outliers mining goal is to find and most other object different objects.Spatio-temporal data-mining refers to extract from the spatio-temporal databases of interest to users with characteristics, spacetime space-time model with the spatio-temporal data universal relations and other some implicit in the database knowledge of the process. The spatio-temporal data both spatial data, and the characteristics of the characteristics of time-series data. The application of spatio-temporal data mining technology and, on the other hand, can make the spatio-temporal query and analysis technology improve the new stage to find knowledge; on the other hand, find knowledge may constitute a knowledge base to establish intelligent GIS system, make GIS become true intelligent information system.Data mining development so far, have developed many data mining platform, such as New Zealand Waikato university development Weka data mining software etc.Based on the above analysis, this article first introduces GML relevant knowledge, data mining, text mining and XML data mining, spatio-temporal data concepts, space time data model, outliers based on data mining, the semi-structured spatio-temporal data GML characteristics, this paper presents an improved GML spatio-temporal data mining algorithm outliers -- TSDBSCAN and TSSOD. Adopt AO second development technology and vs2008C #.net technology, using the MapControl realized the visualization and GML document GML loading. In trials choose two a GML data sources, the experimental results show that when the spatio-temporal data mining analysis of, these two kinds of outliers mining algorithm has very good practicability, efficiency and expansibility. Effective extracting the implicit in GML spatio-temporal data document information. Solved using traditional outliers mining algorithm for mining spatio-temporal data-mining existing problem, achieved GML spatio-temporal data outliers mining. Finally developed a simple GML spatio-temporal data outliers mining prototype system. |
