REODM: Identify Local Outliers in Big Data

Outlier detection is now widely used in various fields. It attracts more and more interests in research. The density based outlier detection methods and the distance based outlier detection methods are the most frequently used outlier detection methods. In big data, the size and dimensions of data is very large. Those features make the conventional methods not suitable for big data. According to the features of big data, we propose a novel REODM method. Based on the rough set theory, we propose a relevant set, which is used to replace the universe set to improve the calculation speed. The REODM method is based on the ensemble learning. It is robust and accurate. We compare the performance of REODM, LOF and KNN methods. The experimental results show that the recall rate and precision of REODM is much better than LOF and KNN method.