%A XIN Liling, HE Wei, YU Jian, JIA Caiyan %T An outlier detection algorithm based on density difference %0 Journal Article %D 2015 %J Journal of Shandong University(Engineering Science) %R 10.6040/j.issn.1672-3961.1.2014.182 %P 7-14 %V 45 %N 3 %U {http://gxbwk.njournal.sdu.edu.cn/CN/abstract/article_1556.shtml} %8 2015-06-20 %X An improved algorithm IMMOD was proposed based on the algorithm MMOD, which considered the difference among different attributes and improved the accuracy of detection. The algorithm introduced entropy to confirm the significance of attribute. The weight of attribute determined by the entropy was used to calculate the weighted distance between objects. In addition, determining and reducing the secondary attributes could guarantee the computational complexity and improve the precision on the high dimensional datasets. The theoretical analysis and the empirical study both showed that the IMMOD could be applied on high dimensional datasets well with a few of parameters and high accuracy, which was better than other algorithms.