Identify subjective news with Pascal
Powered by Recurrent Neural Networks
Fake News Identification uses a pre-trained neural network (specific: Long Short Term Memory) to identify patterns in given news articles. These patterns cumulative produce an output that is either labeled as objective news or subjective news.
Accuracy above 80%
Please take in consideration that the given results are accurate at a percentage of 81% on our evaluation dataset. Therefore, use it as additional feature for your workflow but do not solely rely on the given results.
Use of the Fake News Identification module
Fake News Identification is available within the World News canvas in the CDDmatics Pascal application. The feature is situated under the "Assistent" header. Please note: the system currently only accepts batches up to 16 ("sixteen") news articles. It may take up to several minutes for analysis depending on the size of the articles and the current work load.