This paper introduces how ClaimBuster, a fact-checking platform, uses natural language processing and supervised learning to detect important factual claims in political discourses. The claim spotting model is built using a human-labeled dataset of check-worthy factual claims from the U.S. general election debate transcripts. The paper explains the architecture and the components of the system and the evaluation of the model. It presents a case study of how ClaimBuster live covers the 2016 U.S. presidential election debates and monitors social media and Australian Hansard for factual claims. It also describes the current status and the long-term goals of ClaimBuster as we keep developing and expanding it.

Our society is struggling with an unprecedented amount of falsehoods, hyperboles, and half-truths. Politicians and organizations repeatedly make the same false claims. Fake news floods the cyberspace and even allegedly influenced the 2016 election. In fighting false information, the number of active fact-checking organizations has grown from 44 in 2014 to 114 in early 2017. 1 Fact-checkers vet claims by investigating relevant data and documents and publish their verdicts. For instance, PolitiFact.com, one of the earliest and most popular fact-checking projects, gives factual claims truthfulness ratings such as True, Mostly True, Half true, Mostly False, False, and even "Pants on Fire". In the U.S., the election year made fact-checking a part of household terminology. For example, during the first presidential debate on September 26, 2016, NPR.org's live fact-checking website drew 7.4 million page views and delivered its biggest traffic day ever.

Public figures such as politicians make claims about "facts" all the time. Journalists and citizens spend a good amount of time checking the veracity of such claims. Toward automatic fact checking, we developed tools to find check-worthy factual claims from natural language sentences. Specifically, we prepared a U.S. presidential debate dataset and built classification models to distinguish check-worthy factual claims from non-factual claims and unimportant factual claims. We also identified the most-effective features based on their impact on the classification models' accuracy.