Skip to main content

Table 4 Statistical significance of accuracy differences between xCAPT5 and other models across three datasets

From: xCAPT5: protein–protein interaction prediction using deep and wide multi-kernel pooling convolutional neural networks with protein language model

Model

Martin Dataset

Guo Dataset

Pan Dataset

MARPP (2023)

\(6.32 \times 10^{-5}\)

\(9.20 \times 10^{-4}\)

NA

TAGPPI (2022)

NA

\(8.20 \times 10^{-8}\)

NA

HNSPPI (2023)

\(1.81 \times 10^{-8}\)

\(1.55 \times 10^{-5}\)

NA

PIPR (2019)

\(1.81 \times 10^{-8}\)

\(8.90 \times 10^{-7}\)

\(1.23 \times 10^{-5}\)

Graph-BERT (2023)

NA

NA

\(1.23 \times 10^{-5}\)

FSNN-LGBM (2021)

\(1.13 \times 10^{-3}\)

\(1.79 \times 10^{-8}\)

\(7.76 \times 10^{-2}\)

  1. NA denotes that data is not available for the comparison