Hi, I am reading the CHLP paper using RNN model for mortality prediction, and I am rough on the stats so I’m hoping the statisticians on the forum and @jeremy can help me out.
In the result section, they show the AUC curve comparison between rnn, mlp. logregression(LR), " [MLP: 88.8% (p < 0.01), LR: 86.1% (p < 0.001)", what does (p<0.01) mean? Is this the p value? I only have experience using p value in hypothesis testing, but i’m not sure what it means in the context of model evaluation. Does it mean our confident level about MLP accuracy at 88.8%?
Yup it’s a p value. The context is the start of the sentence: “The RNN model yields an AUC of 93.4% which is significantly higher than the comparative models”. So the p values are the significance of the difference in the other model AUCs vs RNN, where the null hypothesis is that the RNN AUC is the same as the other model.
Thank you for your help! Just to let you know that I’m experimenting asking paper-related questions on the forum, so far the experience seems pretty smooth and positive.
