String similarity measures of edit distance, cosine correlation and Dice coefficient are adopted to evaluate machine translation results. Experiment shows that the evaluation method distinguishes well between "go...String similarity measures of edit distance, cosine correlation and Dice coefficient are adopted to evaluate machine translation results. Experiment shows that the evaluation method distinguishes well between "good" and "bad" translations. Another experiment manifests a consistency between human and automatic scorings of 6 general-purpose MT systems. Equational analysis validates the experimental results. Although the data and graphs are very promising, correlation coefficient and significance tests at 0.01 level are made to ensure the reliability of the results. Linear regression is made to map the automatic scoring results to human scorings.展开更多
This paper discusses the approach to the CET （college English test） oral test based on automatic essay marking （AEM）. The reliability and validity arouses a dispute on Internet-based CET （iB-CET） oral test in Ch...This paper discusses the approach to the CET （college English test） oral test based on automatic essay marking （AEM）. The reliability and validity arouses a dispute on Internet-based CET （iB-CET） oral test in China, leading to the penetration into oral test contents in dialogue, description or comment, and question answering. Then, a probe into transformation from a spoken document into a textual document is touched on a conversion ratio of 72% adopted at the present time, AEM to be carried forward being a necessary hypothesis for Afterwards, this paper focuses on the pipeline in AEM with the content features and the language itself, which is established in two different models in marking for extraction and detection. Through experiments, this paper reveals the fact that an iB-CET oral test can be performed perfectly in the pipeline of spoken documents transformed into textual documents which can be automatically marked. Hence the design of the iB-CET oral test reaches its reliability and validity.展开更多
文摘String similarity measures of edit distance, cosine correlation and Dice coefficient are adopted to evaluate machine translation results. Experiment shows that the evaluation method distinguishes well between "good" and "bad" translations. Another experiment manifests a consistency between human and automatic scorings of 6 general-purpose MT systems. Equational analysis validates the experimental results. Although the data and graphs are very promising, correlation coefficient and significance tests at 0.01 level are made to ensure the reliability of the results. Linear regression is made to map the automatic scoring results to human scorings.
基金The Humanities and Social Sciences Project granted from the Ministry of Education in China(No.10YJA740061,11YJA740121)the Teaching Reform Key Project under Educational Section of Jiangsu Province(No.2011JSJG453)+3 种基金Southeast University Teaching Reform Project for Graduates(No.2010-54)Southeast University Teaching Reform Project for Post-Graduates(No.KJGKT12-06)the Natural Science Foundation of Jiangsu Province(No.BK2008354)the National Natural Science Foundation of China(No.6550181821,60975017)
文摘This paper discusses the approach to the CET （college English test） oral test based on automatic essay marking （AEM）. The reliability and validity arouses a dispute on Internet-based CET （iB-CET） oral test in China, leading to the penetration into oral test contents in dialogue, description or comment, and question answering. Then, a probe into transformation from a spoken document into a textual document is touched on a conversion ratio of 72% adopted at the present time, AEM to be carried forward being a necessary hypothesis for Afterwards, this paper focuses on the pipeline in AEM with the content features and the language itself, which is established in two different models in marking for extraction and detection. Through experiments, this paper reveals the fact that an iB-CET oral test can be performed perfectly in the pipeline of spoken documents transformed into textual documents which can be automatically marked. Hence the design of the iB-CET oral test reaches its reliability and validity.
