Fuzzy Bayesian Classification for Spatial Data Streams with p-trees

Abstract: Enormous amount of geographic data have been and continuously being collected with the advent of modern data acquisition systems like remote sensing, Global positioning system etc. To efficiently process this data, there is a great need to extract the hidden knowledge from these spatial data streams which are unpredictably large in size, complexity and dimensionality. To address these challenges, spatial data mining and geographic knowledge discovery has emerged as an active research field, focusing on the development of theory, methodology, and practice for the extraction of knowledge and useful information from massive and highly complex spatial data streams. This paper emphasizes on Bayesian classification for spatial data. Bayesian is combined with a new data structure called peano count tree for compressing the spatial data, enhancing the scalability and reducing the classifier build time. A new technique called Fuzzy Bayesian is introduced which dramatically increased the performance of the Bayesian classifier. It is ascertained that the accuracy has been improved by Fuzzy Bayesian method with P-trees.

Keywords: Spatial data, Raster format, Bayesian classification, Fuzzy Bayesian classification, Peano Count Tree, bit sequential format (bSQ), Band sequential format(BSQ).