Research on Outlier Mining Algorithm Oriented to Traffic Information
This paper designs a new outlier mining algorithm based on distance through introducing the "key attribute", which reduces the amount of data mining and increases the efficiency of outlier mining, simultaneously, improves the common distance measure and calculates with the improved Modifing Weighted Manhattan Distance(MWMD), the improved mining algorithm cancels the requirements of the parameter setting in the case of without affecting the mining results and gives the isolation degree of outliers.
Jianzhong Wang and Jingang Qi
W. G. Xie et al., "Research on Outlier Mining Algorithm Oriented to Traffic Information", Advanced Materials Research, Vols. 299-300, pp. 1312-1315, 2011