Modern Machine Learning Techniques for Univariate Tunnel Settlement Forecasting: A Comparative Study

Figure 10

Prediction result of surface point number 554 of the metro train line 3 construction in Ningbo city. The surface point goes up in the first phase of the construction and drops down in the later phase due to the underground human interferes. The irregular movements increase the difficulty for forecasting. However, the proposed method can still well predict the surface point movement due to proper rolling window size settings.