Bernstein et al. [9] developed a novel algorithm for discovering topics in short status updates powered by linguistic syntactic transformation, and then the tweets can be grouped into topics mentioned explicitly or implicitly.
Since we were focusing on discovering topics from daily MBPs with statistical analysis, rather than on a specific user or a specific topic, we used the public-timeline interface [3] to collect MBPs (see Figure 1 for an example of such interface).
An example of the discovered topic community using CPM is show in Fig. 11.
Specifically, W reflects the frequency of different words in each discovered topic, while H reflects the topic mix present in each document.
Next, we used computational techniques in natural language processing to cluster the abstracts into neuroscience topics and studied their dynamics and concordance of these discovered topic clusters with the thematic organization provided by the SFN.
4 Topic correlation Calculating correlations between discovered topics.
The technique has been employed both to discover topics in text collections and to structure document sets for advanced searching.
In this paper, we propose to use sentence-level association rule mining to discover topics from documents.
The goal is to help users discover topics that they'll be interested in quickly, and then foster productive conversation.
