automated machine learning

The datasets used for TKDE manuscript. These datasets are widely recognized and validated within the summarization research community, and using these well-established datasets ensures that our results are comparable with the vast majority of related work.In each dataset, we use 80% as the training set, while the remaining 20% as the test data. The range of datasets used for collecting knowledge is also consistent with the above. For the large dataset, we only use a part of it because collecting knowledge is time-consuming.