Abstract:
Expertise is the important resource in software development. It is difficult to find skills of experts in an organization. Normally, administrator identifies skills with personal experiences with the experts. The purposes of this study are: 1) to develop and evaluate a knowledge base system that extracts and classifies knowledge from software archives, and, 2) to develop data representation of the knowledge base in supporting and identifying domain experts and evaluate the effectiveness of the developed data representation. The developed knowledge base system extracts expert skills from the software archives and classifies the skills based on software concept ontology which derived from Open Project
Directory (OPD). The skills are categorized by using similarity measurement, the Jaccard coefficient. The accuracy of knowledge classification is evaluated by using precision and
recall. Knowledge base data representations were developed using spreadsheet, and visualization tools: TreeView and TreeMap with Node Graph.
Evaluating the effectiveness of data representations was achieved through comparing the time used in completing tasks with different complexity, and, users’ satisfactions. TreeMap with
Node Graph spent least time in showing pairs of experts and skills, and, also shows the positive changes of users’ behaviors in identifying experts and skills, therefore, has received the highest results of users’ satisfaction.