Developing Dataset Management of Speech Recognition Based Automatic English-Speaking Skill Testing System

Abstract

Speech recognition technology has been widely used to aid English language testing, such as for pronunciation and speaking practice for various use cases including job interview simulation. We developed an app, powered by speech recognition technology and regex text processing, that automatically score participants’ speaking performance. However, this application lacks dataset management for the speaking test maker to manage the indicators of accepted answers. A feature to effectively manage the dataset and efficiently ease the speaking test creation is needed. This paper presents the development of dataset management feature to support the speech recognition based automatic English-speaking skill testing system. All elements that support the feature in development were tested through Blackbox testing with various scenarios. The elements were proven to work as intended to support the existing system. The new feature has brought efficiency for the English-speaking test maker.