Information about the state and planning of the speaker is obscured in traditional classifications of disfluencies which are generally at the word level. This study delves into the acoustic and prosodic information of repetitions, one of the most common disfluencies. A hierarchical clustering of prosodic features reveals three subsets of repetitions, each reflecting different problems in planning.

Utilizing context information—e.g. location, user aspects, or hardware capabilities—enables the presented generic framework to automatically control the selection and configuration of visualization techniques and therewith provide interactive illustrations, displayed on small mobile devices. For contextannotated data, provided by an underlying context-aware world model, the proposed system dete...

We create a weighted lexical network derived from the cosine similarities of financial news feeds to compare two clustering methods, Newman's Modularity method and hierarchical clustering. We find that hierarchical clustering, clustering documents according to shared unique terms, shows results that are closer to expectation.