A Recommender System for Technical Software Documentation in Wikis

JORGE ANDRES DIAZ PACE; MATIAS NICOLETTI; SILVIA N. SCHIAFFINO; RODRIGO CRISANTO; MARTIN NAJLE

Bariloche

Congreso; Congreso Bienal de IEEE Argentina - ARGENCON 2014; 2014

IEEE Argentina

Technical knowledge is an important asset of today?s software development projects. Collaborative and content management environments, such as Wikis, are becoming common nowadays for capturing such knowledge and sharing it among the project stakeholders. However, different stakeholders are often interested in subsets of the whole document repository, depending on specific information needs or tasks carried out. Thus, a common problem is that of information overload, which usually makes stakeholders unhappy when they are not able to find relevant documents. In this context, we have developed a recommendation system called SODER, which integrates a Wiki platform with a chat mechanism, and applies user profiling techniques in order to automatically infer the interests of the stakeholders as they work with the Wiki. A novel aspect of SODER is its ability to merge information from different sources when building the profiles. A preliminary evaluation of SODER has shown its potential for recommending relevant technical documents.