MOOD - News AMR dataset - Hackathon 2022

This dataset has been collected from four Epidemiological Surveillance Systems (EBS) to be used in an hackathon dedicated to AMR (antimicrobial resistance) for the MOOD summer school in June 2022. The choosen EBS sources are ProMED, PADI-web, Healthmap and MedISys. The collected data are news dealing with epidemiological information or event. This dataset is composed of 4 sub-datasets for each chosen EBS. Each sub-dataset is annotated according to 3 main classes (New Information, General Information, Not Relevant). For each news labeled as New Information or General Information, another annotation is provided concerning host classification with 7 classes (Humans, Human-animal, Animals, Human-food, Food, Environment, and All). This second annotation provided 4 sub-datasets. The aim of the annotation task is to recognize epidemiological information related to AMR. An annotation guideline is provided in order to ensure an unified annotation and to help the annotators. This dataset can be used to train or evaluate classification approaches to automatically identify text on AMR events and types of AMR issues (e.g. animal, food, etc.) in unstructured data (e.g. news, tweets) and classify these events by relevance for epidemic intelligence purposes.