Abstract
With the advent of the World Wide Web, availability of data is no longer an issue. However organizing and analyzing information is becoming an increasingly complicated task. Formal representation and structuring of knowledge is necessary for software systems to utilize knowledge. Visual Knowledge Builder is a system that actively supports incremental formalization during information capture and analysis. Though the current VKB version supports incremental formalization of knowledge, the onus of actually carrying out the formalization is on the user. The user has to carry out the formalization of visual workspace without any input/suggestions from the system. This can be a difficult and tedious task if the size of the workspace is very large. My thesis involves the use of spatial, textual and temporal attributes of the workspace by VKB to provide feedback to the user regarding formalization of the workspace. This feedback will be provided using a suggestion mechanism. As the users continue their work, workspace analysis occurs in the background. When a candidate formal structure is identified, it is represented as a suggestion to the user. I explore different ways of using spatial, textual and temporal information for making a variety of suggestions to the user. I also investigate the premise that use of a combination of the above attribute types for recognizing structure in informal representation gives rise to better quality suggestions than using these attribute types individually.
Maloor, Preetam (2001). Combining spatial, textual and temporal analysis for the construction and incremental formalization of knowledge. Master's thesis, Texas A&M University. Available electronically from
https : / /hdl .handle .net /1969 .1 /ETD -TAMU -2001 -THESIS -M347.