An FCM-based Personalized Affective Model for Agile Software Development

Abstract

Abstract-Developers' emotional stability in agile software development (ASD) teams is an important factor affecting the success of a project. Traditional Happiness Chart depends on people's self-report and is unable to produce predictive analysis to facilitate decision-making. In this paper, we proposed a Fuzzy Cognitive Map (FCM) based method for analyzing ASD developers' emotional stability. We designed an easy-touse survey questionnaire which can help establish the values of the causal relationship weights to support personalization of the proposed FCM model. The FCM affective model produces two metrics for assessing a given user's emotional stability: 1) the stable state mood value, and 2) the number of iteration steps for the FCM model to reach equilibrium. The two metrics are evaluated through a real-world experiment involving 20 undergraduate students. The results show that they are strong explanation factors for a user's emotional stability with 99% confidence. A regression model based on the user study has also been presented in the paper to help researchers study the relationships between the two metrics and a person's emotional stability computationally in the future.