Fuzzy Co - The Location Pattern Mining

In recent years, with the deepening of research and application of the classical collection, the shortcomings of the classical collection are also increasingly exposed such as loss of information, the border is too obvious. Therefore it needs to find a new method to deal with data of fuzziness and improve the classic collection. Fuzzy set which is well extending the classical collection was generated in this background. It could solute the common fuzzy phenomenon more scientific than the classical collection in the objective world. On the other hand, along with the rapid development and progress in data gathering, data processing and Internet, spatial data research gradually deepened. As one of the important research in spatial data mining. Co-Location spatial Pattern Mining aimed at finding the feature whose instance is frequent co-locate in neighboring domain. In addition to the spatial location information, spatial data usually also contains the attribute information, so discuss ambiguous Co-Location model mining has a very important significance.First, we analyze current situation of the study. This paper describes the basic definition, the related work and major challenges of Co-Location rule mining. We also introduce fuzzy set theory and its application to association rules mining.Secondly, we define the attributes discrete process for the spatial data with quantitative attributes. Then, we redefine the issue and propose fuzzy Co-Location pattern mining algorithms. We also analyze the properties and time complexity of the algorithm.Third, according to observation and study we proposed two pruning methods after the analysis of the fuzzy Co-Location rule mining algorithms. Through experiments on synthetic data, it is proved that the pruning strategy is effective. Experiments on actual data show that our study over fuzzy Co-the Location Pattern Mining is significant.At the last, the conclusion and future work were presented.