Automatic Evaluation of Communication Competency in Diverse Environments

Abstract: Effective communication is a critical social skill that helps us understand and relate to others. It is also crucial in job-related interviews. In this project, candidates' communication skills in two types of behavioural interviews-a written interview that includes a brief essay and an interface-based, asynchronous video interview-are methodically studied and automatically measured. Next, by utilizing deep learning methods and machine learning XGBOOST, we suggest a prediction model that makes use of automatically extracted multimodel features such as audio, visual, and lexical. While all currently available technologies predict essays with an accuracy of 80-90%, our XGBOOST technology predicts essays with an accuracy of 95-96%. The capacity to effectively and efficiently convey knowledge to others is known as communication proficiency.

Keywords: Communication Skills, XGBOOST, Interviews, Deep Learning, CNN.