Threshold Regression Forecast Model for Dam Safety Monitoring and its Application
Affected by both of the internal and external factors during work process, the dam monitoring effect variable often generates complex cycles such as jumps, compendium, and re-jumps, which shows a high degree of non-linearity that brings a lot of trouble in the establishment of regular dam safety monitoring forecast model. Drawing on the threshold thought of the threshold autoregressive model, the threshold factor and the threshold value are introduced in the conventional regression model. By making use of the suggested best fitting correlation analysis method to determine the threshold factor and the threshold value, the threshold regression model is established. Application example shows that the method for dam safety monitoring forecast can obtain satisfactory results.
Yusaku Fuji and Koichi Maru
X. F. He et al., "Threshold Regression Forecast Model for Dam Safety Monitoring and its Application", Applied Mechanics and Materials, Vol. 36, pp. 182-186, 2010