Hydrological data anomaly detection method based on spatio-temporal information

Abstract

The invention discloses a hydrological data anomaly detection method based on spatio-temporal information. The method comprises the following steps: dividing associated sites; dividing a water level time sequence; obtaining a model output result by using the trained convolutional neural network (CNN) model, carrying out residual prediction on the model output result by using a Markov chain (MC), and judging an abnormal station according to the model output result and the predicted residual; obtaining abnormal conditions of the to-be-detected station and all associated stations; and performingresult fusion by adopting a dynamic distribution DS evidence theory (DADS) algorithm to obtain a hydrological data exception prediction result. According to the method, the influence of rainstorm seasons on hydrological data is fully considered, the detection precision is improved, a shuffled frog leaping algorithm (SFLA) is introduced to improve convolutional network parameters, an MC algorithm is added to carry out residual prediction, and the accuracy of prediction data is improved; and finally, through a dynamic distribution D-S evidence theory, fully considering spatial factors, and fusing multi-associated site prediction results, so the false alarm frequency is effectively reduced.

Description

technical field [0001] The invention belongs to the field of data mining, and relates to a data anomaly detection method, in particular to a hydrological data anomaly detection method based on spatio-temporal information. Background technique [0002] In recent years, due to deforestation and predatory use of forest resources, the vegetation in the Yangtze River and Yellow River basins has been damaged, the land has become desertified, and floods have occurred from time to time. According to incomplete statistics, since the founding of the People's Republic of China, the average annual flood-affected area in my country is 134 million mu, and the disaster area is 76 million mu, with direct economic losses reaching tens of billions of yuan. In view of the above situation, how to use effective methods to accurately and quickly forecast floods is of great significance for flood control and disaster reduction, ecological balance adjustment, and regional water resource scheduling....

AI Technical Summary

Problems solved by technology

However, the existing anomaly detection algorithms are not enough to deal with the complex characteristics of hydrological data such as seasonality, randomness, and time-space correlation. Therefore, there is still a large room for improvement in the accuracy of anomaly detection, which is worth investing a lot of time. time and effort to conduct research

Images