Analysis and Evaluation of Individual Performance in Team Task

Abstract

Opinion expression in team task plays an essential role in individual performance evaluation. So this study set out to answer the question: How to evaluate individual performance mostly from interaction standpoint. The goals are twofold: 1. to determine whether the three aspects of communication skills, namely “how to speak”, “what to say” and “who to communicate”, are effective features for evaluating individual performance; 2. to design a methodology evaluating individual performance by computation. For the former, the effectiveness of features in speaking cues, communication patterns and speech content are evaluated based on Team Cooperation Database. As for the latter, the evaluation is implemented utilizing Support Vector Machines with mean accuracy 88% which outperforms other six machine learning models, significantly improving accuracy over the baseline. Experimental results indicate that communication patterns and speaking cues are major factors affecting individual performance. Additionally, it can also be concluded that individuals who speak after listening the opinions of others and share their ideas with teammates widely and fully can gain more acceptance in performance evaluation.