Abstract:
This thesis presents an ontological approach to emotion analysis, for software development
teams. Software development teams, especially for open-source projects, are usually
geographically dispersed. Possibility of arousal conflicts becomes a norm; not an exception. In the
absence of human contact, it becomes quite challenging to virtually facilitate.
Emotional state of team members affects the performance and outcomes of the team
processes. Therefore, emotion analysis becomes an important area for investigation. In order to
study emotion analysis, it is intuitive to first understand what emotions are. Emotions are usually
described as one’s feelings towards an entity. However, there is no universally accepted definition
available for emotions.
Several scholars have discussed emotions from different perspectives. This has led to
introduction of various theories and models for describing emotions. This study presents an
ontology of emotions, built on one such model, for performing emotion analysis. An analyzer is
developed using this ontology, which performed analysis on emails taken from mailing lists of
Apache Software Foundation projects. The analyzer is able to recognize emotions with an accuracy
of 61.3%.
The results indicate that software developers do express emotions in their discussions.
These emotions can be identified with the help of an application, similar to the analyzer. Thus, the
emotion analysis can be used to study the emotional state of teams, and hence, can act as an aid in
team facilitation.
