Date of this Version
In this study, we demonstrate that a visualization model can determine the effectiveness of searching a bibliographic database, when three descriptive metadata fields are compared. The Inspec database was searched to create a bibliography of articles about a broad scientific topic, interplanetary travel. We collected metadata from 823 Inspec articles and used the Sci2 Tool to create co-occurrence networks based on subject terms, title keywords, and classification codes from each of the articles. The Watts-Strogatz clustering coefficient model was used to create molecular organization of the networks. This method identified subject domain clusters for each of the three selected metadata elements and subject classification codes were extracted from all the clusters obtained. All data obtained was converted into a common metadata element (classification codes), allowing for the comparison of data from the initial search and from all the subject clusters identified in the visualization process. A set of eight subject codes were found to describe the Main Subject Domain of interplanetary travel. The results also show that searching with classification codes produced the best outcome, the second best option is using subject terms, and the least effective search technique is using title keywords. These results, using visualization, corroborate previous studies.