Multi-verifier: A novel method for fact statement verification

摘要

Untruthful information spreads on the web, which may mislead users or have a negative impact on user experience. In this paper, we propose a method, called Multi-verifier, to determine the truthfulness of a fact statement. The basic idea is that whether a fact statement is truthful or not depends on the information related to the fact statement. We utilize a popular search engine to collect the top-K search results that are most related to the target fact statement. We then propose support score to measure the extent to which a search result supports the fact statement based on semantic constituent analysis, and we make use of credibility ranking to rank the search results. At last, we combine the support score and credibility ranking value of a search result to evaluate its contribution to a fact statement determination. Based on the contributions of the search results, we determine the fact statement. Our proposals are evaluated by experiments and results show availability and high precision of the proposed method.