13 commits | Last commit ≈ 86 months ago
An application that uses cross-perspective topic modeling to extract topics and opinions from text and provides insight into how they change over time.
cptm is an implementation of cross-perspective topic modeling, which is is a topic modeling method that simultaneously extract topics and opinions from text. Topics are learned from nouns and opinions from adjectives. To use this tool, you need a corpus that is divided into perspectives. In the DiLiPaD project a perspective refers to a political party. We used this tool to extract topics and opinions from parliamentary proceedings.
A new approach to the history of parliamentary communication and discourse