Academic Literature Knowledge Q&A Dataset

This dataset is the supporting data for the paper "An intelligent consulting Q&A system based on machine reading comprehension", including the following four files:1. Self-built academic literature knowledge Q&A dataset. The extracted machine reading comprehension dataset is obtained by manual extraction + data augmentation with selected medical literature as the data source.2. Deep learning training result statistics table. Deep learning fine-tuning of training data using multiple models and multiple schemes and analysis and comparison of training results.3. Multi-indicator machine learning dataset for paragraph recall. The multi-metrics are combined to build multiple pairs of "question-passage" data, and the positive sample is the one that matches the correct passage, and the negative sample is the opposite.4. Paragraph recall indicator test table. After machine learning, we obtain the weights of each indicator and take 200 items of each of the four types of questions as test data, and test the recall accuracy of the data under the new weights.