Outlier Detection of Hydrological Time Series Based on ARIMA-SVR Model
SUN Jianshu
LOU Yuansheng
CHEN Yujun
Abstract：Due to weather,terrain and other complex factors,hydrological data usually have many unusual values,and it has an important impact on decision-making.In this paper,an anomaly detection algorithm based on ARIMA-SVR is proposed,and it improves the quality of anomaly detection.First,the ARIMA model is used to predict the linear autocorrelation of the hydrological time series,and then the SVR-model is used to predict the nonlinear part.Secondly,the predicted results are summed up and the confidence intervals of confidence P is obtained.At last,the actual value which is not within the confidence interval is an outlier. We use The Liuhe hydrological station measured data is used to verify,and experimental results show that the proposed algorithm can effectively detect the outliers in the hydrological time series,while the specificity and sensitivity are maintained at a high level.
Keywords：hydrological time seriesanomaly detectionARIMA-modelSVR
Publication Date：2018-01-01
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：6( 225-230 )