Summary: | GBLENDER is an intuitive user-friendly visual interface to formulate and process queries in graph
databases using a frequent-structure based approach. As the application is not released on a public
domain, there is a lack of users available to evaluate the performance of the application. To address
this problem, this project proposed the development and implementation of a simulator that applies
concepts of Human Computer Interaction to visually formulate and process queries on GBLENDER. As
GBLENDER seeks to exploit the latency between query formulations by prefetching graph results and
pruning false results, analysis of the predicted user time and the prefetch time would provide useful
information on the performance of GBLENDER. We have included various modifiable settings like
adjusting the database indexing settings, user skills level and environment to examine the
performance at different levels. The raw results collected were displayed in many types of charts,
making it digestible for researchers to quickly locate certain data.
|