Assessment of information credibility in Twitter
Thompson, Kyle Maurice
MetadataShow full item record
In today’s society, social media has become one of the primary sources of obtaining information. Citizens utilize social media platforms to familiarize themselves with and learn about, current events around the world. There is an increase in individuals that depend on social media to be valid and provide correct information on newsworthy topics. This research provides a unique automated approach that incorporates human intelligence and text mining techniques. The phrase “What vs. Who” is used to separate the past research from this approach. The automated approach includes techniques similar to text mining which uses Python’s Natural Language Toolkit to incorporate natural language processing and search through individual tweets for certain keywords. To provide credible information, news articles, verified Twitter accounts, and criminal and domestic release records, which possess official power, were used as the golden standard and deemed credible. Crowd sourcing was then used to compare results with the automated approach.