RAINWATER FLOW-IN AMOUNT FORECASTING DEVICE APPLYING NEURAL NETWORK

摘要

PURPOSE:To automatically and accurately forecast the peak rainwater flow-in amount by storing the learned weight factor of the neural network in the back propagation method and forecasting the peak rainwater flow-in amount while using the weight factor based on the rainfall intensity data for each day. CONSTITUTION:A neural network learning device 45 learns the weight factor of the neural network by the back propagation method based on the rainfall intensity data being the learning input data stored in a learning input pattern preserving device 42 and the peak rainwater flow-in amount being the learning output data stored in a learning output pattern preserving device 44. A learned weight factor preserving device 46 stores the weight factor learned by the neural network learning device 45. Further, a neural network forecasting device 47 forecasts the peak rainwater flow-in amount by the neural network based on the rainfall intensity data for each rainy day to be inputted by a mesh rainfall intensity arithmetic device 41 using the weight factor.