Dynamic outliers data identifying and detection based on chaotic
Jianzhou Wang · Journal of Tsinghua University(Science and Technology) · 2005
Identifying outlier is an important data analysis function. But people often try to minimize the influence of outliers or eliminate them all together in traditional outlier data analysis. This can result in the loss of important hidden information. This paper proposes a new method for outlier detection and mining. The slope was computed between two data points first in time series, and was compared with chaotic forecasting slope to detect data deviate set. Dynamic sum square minimum error detection was used to deviate set and confirm outlier set. The masking effect is solved and the shortcoming is overcome that outlier data detection can not be measured in outlier data analysis.