Automatic Scoring on English Passage Reading Quality

Junbo Zhang, Fuping Pan, Yonghong Yan

Abstract: In this paper, the computer automatic scoring for English discourse oral reading was studied. We analyzed the oral reading voices with speech recognition technology, and extracted a series of features in terms of pronunciation and fluency, and then mapped these features to scores. In the testing of English discourse oral reading for 4000 middle school students, the average scoring difference between machine and human teacher was 0.66, while the scoring difference in human teachers was 0.57. The experience result shows that this system can be used in practice.