Machine-learning based automatic assessment of communication in interpreting

Communication assessment in interpreting has developed into an area with new models and continues to receive growing attention recent years. The process refers the of messages composed both “verbal” “nonverbal” signals. A few relevant studies revolving around automatic scoring investigated fluency based on objective temporal measures, correlation between machine translation metrics human scores. There is no research exploring machine-learning-based in-depth integrating parameters delivery information. What remains fundamentally challenging demonstrate which parameters, extracted through methodology, predict more reliable results. This study presents original aim propose test a learning approach automatically assess communication English/Chinese interpreting. It proposes build predictive using algorithms, extracting for delivery, applying quality estimation model information describe final model. employs K-nearest neighbour algorithm support vector further analysis. found that best machine-learning built all features by Support Vector Machine shows accuracy 62.96%, better than 55.56%. results pass level can be accurately predicted, indicates are able screen interpretations exam. first supervised fidelity point great potential little evaluation involved process. Automatic expected complete multi-tasks within brief period taking holistic analytical approaches accuracy, delivery. proposed system might facilitate human-machine collaboration future. generate instant feedback students evaluating input renditions or abridge workload educators education screening performance subsequent scoring.