An Experiment in Software Development Risk Information

The following report summarizes the results of an experiment that uses terminological structures derived from the application of knowledge summarization, analysis, and visualization (K-SAV) technology to textual data from the Software Engineering Risk Repository (SERR) resident at the Software Engineering Institute. This study evaluates the use of several tools including shared word clustering [Monarch 94] and a co-word analysis software program, leximappe [Teil 92]. The experiment seeks to determine whether an application of co-word analysis to baseline risk assessment data would enable a reduction of the information load while simultaneously providing a succinct but encompassing picture of the risk information within the program. This study is based upon a somewhat limited data set. Nevertheless, the results of this investigation are encouraging and suggest that there may be value and potential for the effective use of co-word analysis and K-SAV technology more generally in risk management. Additional investigations are underway to confirm, alter, or challenge the results.