Mining Spatial Rough Set Theory Co-Location Mode

Rough set theory is a theory and tool to manage uncertain knowledge. It especially suitable for expressing, learning and summarizing the knowledge that is uncertain, incomplete and noisy. Its effectiveness has been successfully applied in many science and engineering fields. Rough set theory can deal with imperfect data, which stressed that the data cannot be distinguished, imprecise and ambiguous. On the other hand, with the rapid development and progress of the network technology and spatial data collection technology, the spatial data is complex and volatile, people also want to discover knowledge from the spatial database. An emerging field of research-spatial data mining arises. Spatial data mining combines data mining and spatial database technology. Space Co-Location Pattern Mining is a special case of spatial association mining for discovering frequent juxtaposition of space events or feature subset. Spatial data usually contains multiple non-spatial attributes in addition to the space attribute information, so it is very meaningful that rough set theory is applied on the space Co-Location Pattern Mining. The paper tends to apply rough set theory to the space Co-Location Pattern Mining model in order to solve multiple non-spatial attributes.First of all, state research status, basic definitions and related work of space Co-Location Pattern Mining Research, and give a brief introduction to the process and algorithm of mining on the classic Space Co-Location mode.Secondly, introduce the basic concepts of rough set theory, application of rough set theory in the association rule mining and list a series of related work on mining of rough association rules.Third, study space co-location mode mining based on rough set theory, come up with related definitions, algorithms and pruning strategy and conduct the experimental analysis on time performance and mining results of the algorithm.Fourth, verify the feasibility and effectiveness of the algorithm through a large number of experiments with simulated data and real data. Finally, sum up the work and look far ahead into the future research on spatial co-location model mining based on rough set theory.