Skip to main content

Table 1 Best performing FFNN sorted by \(R^2(\mathrm {D}^\mathrm {T})\) after Phase 2

From: Application of artificial neural networks for the prediction of interface mechanics: a study on grain boundary constitutive behavior

ID

Architecture

Phase 2

  

\(\varvec{\{N,n,a(x)\}}\)

\(\varvec{R^2(\mathrm {D}^\mathrm {T})}\)

\(\varvec{R^2(\mathrm {D}^\mathrm {C})}\)

\(\varvec{R^2(\mathrm {D}^\mathrm {V})}\)

FFNN1

\(\{4,8,\tanh (x)\}\)

0.9475

0.9756

0.9756

FFNN2

\(\{4,7,\tanh (x)\}\)

0.9442

0.9736

0.9736

FFNN3

\(\{3,8,\tanh (x)\}\)

0.9401

0.9727

0.9727