Multilingual Rumour veracity classifier
User generated content such as tweets often make claims that are unsubstantiated and possibly untrue. This service attempts to classify whether a text is discussing a rumour that is likely to be true, likely to be false, or if the rumour is unverified or the classification is unclear. The classifier is based upon Twitter-XLM-RoBERTa finetuned using an upsampled version of the RumourEval shared task 2017 dataset.
Default annotations | |
:Veracity | Annotation spanning the whole text with features "rumour_label" (the raw label "true", "false" or "unverified" from the classifier), "status" (a more human-oriented version of the rumour_label) and "confidence" (the confidence score) |
Use this pipeline
You can process up to 150 documents per day free of charge using the REST API, at an average rate of 2 documents/sec. Higher quotas are available for research users by arrangement, contact us for details.
The API endpoint for this pipeline is:
You can process any amount of data with this pipeline on a pay-as-you-go basis, for £0.80 per hour. This can be data you upload yourself, data you collected from Twitter, or the results of a previous job.