Detecting Knowledge Innovation through Automatic Topic Labeling on Scholar Data
Diogo Nolasco, Jonice Oliveira · 2016
Knowledge innovation can be detected by scientific and technology production. A challenge on knowledge innovation analysis is the identification of knowledge areas in text documents - called topic modeling - and presenting the related concepts in an understandable way. The most used and reliable approaches are based on manual labeling, which is impracticable when we analyze a huge period of time, a large number of documents or different domains. We propose a method of automatic labeling. Evaluation results pointed that the proposed method can effectively generate labels for topics with a similar efficiency of human generated labels.