Classification accuracy for LOO cross-validation combined with different models and arrangements (i.e., whether <i>D</i>_<i>PS</i> or were employed and with which D2V vector size: 32, 64, 128, or 256), considering the PS dataset.

Highlighted in bold is the best result for each configuration.各配置对应的最优结果均以粗体高亮标注。