Abstract:
Most existing outlier detection algorithms ignore local information of data sets, they are of low accuracy. We adopted a twophase algorithm based on k-means clustering algorithm, defined a new local stray factor as the standard to judge whether data objects are outliers. We also improved the process of detecting outliers and solved the above problem. Experiments show that our algorithm overcomes the shortcomings of existing methods, ensure the algorithm has linear time complexity and is able to find outliers in data sets more accurately and effectively.