PREDICTION OF MINING INDUCED LAND SUBSIDENCE USING SUPPORT VECTOR MACHINES

摘要

This study applies Support Vector Machines (SVMs) to model the complex time dependent behavior of mining induced land subsidence. The highly nonlinear mapping relationship between the accumulated subsidence and the subsidence history was regressed from data taken from field measurements. Typical case studies employing monitored subsidence induced by underground mining are discussed and the application results show that SVMs provides a promising alternative for mining subsidence prediction.