[S18] Runoff and Flooding

[127] APPLICATION OF LONG SHORT-TERM MEMORY (LSTM) NETWORKS APPROACH FOR RIVER WATER LEVEL FORECASTING USING MULTIPLE RIVER BASINS: A CASE STUDY FOR SRI LANKA

*Diani ABEYRATHNE¹, Ryo KANEKO², Kei YOSHIMURA³ (1. Graduate School of Engineering, The University of Tokyo, 2. Institute of Industrial Science, The University of Tokyo, 3. Institute of Industrial Science, The University of Tokyo)

Keywords:data-driven methods、feature selection、flood forecasting、Long Short-Term Memory (LSTM)