Outlier Mining Algorithm Based on Data-Partitioning and Density-Grid

Existing outlier mining algorithms such as FOMAUC are based on density-grid. These algorithms have the problems of inefficiency and bad-adaptability for various data sets, so this paper proposes an outlier mining algorithm based on data partitioning and grid-density. Firstly, the technology of data partitioning was applied. Secondly, the nonoutliers were filtered out by cell and the temporary results were saved. Thirdly, the improved CD-Tree was created to maintain the spatial information of the reserved data. After that, the nonoutliers were filtered out by micro-cell and were operated efficiently through two optimization strategies. Finally, followed by mining by data point the resulting outlier set was obtained. Theoretical analysis and the experimental results show that this method is feasible and effective, and that has better scalability for dealing with massive and high dimensional data.