èצӰÏñ

Speaker(s): David Mimno, assistant professor in the Information Science department at Cornell University

In the last ten years we have seen the creation of massive digital text collections, from Twitter feeds to million-book libraries. At the same time, researchers have developed text mining methods that go beyond simple word frequency analysis to uncover thematic patterns. But models are not enough. When we combine big data with powerful algorithms, we can enhance qualitative perspectives with quantitative measurements. But these methods are only useful if we distinguish consistent patterns from random variations. In this talk I will describe my work with Matt Jockers building reliable topic-mining methodologies for humanists, with examples from a corpus of 3000 19th century novels.