Skip to main content

Table 7 Standard deviation of model residuals for various GRNN models implying the initial, actual frequency amplification factors (RP, NP, TP cases, last three columns) for various combinations of site parameters

From: Deriving amplification factors from simple site parameters using generalized regression neural networks: implications for relevant site proxies

Number of parameters Considered site parameters Error (RP) Error (NP) Error (TP)
All (6) Depth + f 0 + V sm + C v +V s30 + V bedrock 0.0011 0.0032 0.0053
3 (best triplet) f 0 + C v  + V s30 0.0079 0.0079 0.0118
2 (best pair) f 0 + C v 0.0251 0.0233 0.0278
2 (convenient pair) f 0 + V s30 0.0782 0.0382 0.0339
1 (best) C v 0.0725 0.0652 0.0622
1 (usual) V s30 0.1038 0.0715 0.0733
1 f 0 0.099 0.0678 0.0563
Overall initial variability term \(\sigma_{\text{m}} \left( \theta \right)\) 0.1178 0.0846 0.0896