SGMss: A Real-time Dynamic Grey Model for Soil Settlement Prediction

Soil settlement, as a general geological disaster in the world, can cause huge damages to the safety of people's lives and property. Accurate prediction of soil settlement can efficiently support disaster prevention. Existing prediction models of soil settlement are mostly based on grey model or machine learning algorithms, however, these prediction models are mostly static, and rarely have the mechanism to dynamic updating the model based on real-time monitoring data, resulting in the old data reduces the prediction accuracy. In this paper, based on the most widely used grey model GM (1,1), combined with the sliding window technology, a prediction model Sliding Grey Model for soil settlement (SGMss) is produced. Experimental results show that SGMss can effectively apply in the soil settlement prediction applications.