Outlier Detection and Correction During the Process of Groundwater Lever Monitoring Base on Pauta Criterion with Self-learning and Smooth Processing
Limin Li, Zong-zhou Wen, Zhongsheng Wang · Communications in computer and information science · 2016
Underground water level monitoring is vital for human sustainable development, accurate understanding of underground water level is vital for government decision-making. But usually data collected by the ground water level monitoring device contains a large number of outliers which diverge from the normal data, and it will affect the actual judgment. In this paper, on the basis of a large number of field data analysis, outliers monitoring methodology was proposed based on self-learning Pauta criterion, and outliers correction method was proposed based on smoothing processing. Pauta criterion can detect outliers under the condition of the confidence probability of 99.7 %, for the abnormal values points detected, through the data statistics before, using mean of several values before and after to obtain the revised data. Experimental results show that outliers of underground water during the process of level monitoring can be effectively detected, and ideal curve was gained through revising.