Fig. 4From: Improving patient self-description in Chinese online consultation using contextual promptsAll experimental results on the three corpora are exhibited in this figure. The whole figure consists of 15 subplots, a–o. Each column of the figure is one data set, each row is one metric. In one subplot, the x-axis is the number of prompts, y-axis is the corresponding metric. For instance, subplot (a) summarizes the accuracy of all methods from 1 to 10 prompts on pediatrics data set. The higher the accuracy, macro-averaged F-score, macro-averaged AUC and macro-averaged MCC the better. The lower the entropy, the betterBack to article page