Table 3 Comparative performance metrics for energy and force FF predictions on the training set.

From: Enhancing ReaxFF for molecular dynamics simulations of lithium-ion batteries: an interactive reparameterization protocol

FF

Energy

Forces

\(\mathbf {R^2}\)

RMSE \((\hbox {eV})\)

\({R^2}\)

RMSE \(\hbox {eV}\) Å\(^{-1}\))

Yun et al.

\(-0.093\)

0.562

0.227

\(5.1\times 10^{-3}\)

Wang et al.

\(-0.093\)

0.562

0.227

\(5.1\times 10^{-3}\)

This work

0.293

0.452

0.377

\(4.6\times 10^{-3}\)

  1. To evaluate the efficacy of the FFs we compute the Coefficient of Determination (\(\textrm{R}^{2}\)) and the Root Mean Square Error (RMSE) metrics.
Back to article page