Wpływ klasyfikacji wzorców na dokładność neuronowej prognozy przekazywania drgań pochodzenia górniczego z gruntu na fundament budynku

Abstract

The paper deals with the prediction of the soil-structure interaction in the case of transmission of ground vibrations from mining tremors to medium-height building foundations. Neural networks were proposed for the prediction of mine-induced ground vibrations transmission to building foundation. The information about mining tremors classification was taken into account in the neural network input vector. The procedure of neural network training patterns reproduction was also proposed in order to obtain better results of neural network approximation.