Replication Data for: Identifying unreported links between ClinicalTrials.gov trial registrations and their published results

This is the dataset we used to train and evaluate for the paper Identifying unreported links between ClinicalTrials.gov trial registrations and their published results. This dataset is collected on 29 September 2020. The corresponding code and the structure of the dataset can be found in https://github.com/evidence-surveillance/unreported_link_identidication. We highly recommend building your own datasets with more specific criteria for specific applications rather than directly applying this dataset. This dataset includes extracted Clinical Trial registrations, extracted PubMed articles' titles and abstracts, the automatically extracted links between clinical trials and PubMed articles (already shuffled), manually curated links and the transformed data with corresponding vectorizers. This dataset contains 27,280 automatically constructed links (used for training and evaluation) and 90 manually curated links (used ​for evaluation only). The first 27,280 links refer to the automatically constructed links and the last 90 links refer to the manually curated links.