摘要
本文分析了自动作文评分系统BETSY的工作原理,并采用987个样本的训练集、200个样本的验证集,实证检验该系统在英语作文自动评分中的效果。统计结果显示,BETSY的评分结果与人工评分结果有较强的相关性。实验同时分析了影响BETSY评分精度的原因。
In the present study,principles of BETSY were analyzed and effects of BETSY in AES were quantitatively studied on the basis of the 987-sample training set and 200-sample validation set. The statistics showed that a strong correlation was observed between human rating and BETSY rating. In addition,factors affecting measure precision and measure results were analyzed.
出处
《当代外语研究》
2011年第1期44-47,62,共4页
Contemporary Foreign Language Studies
基金
2008年度湖北省教育厅高等学校省级教学研究基金项目“大学英语网络题库的建设与应用”(项目编号为2008225)研究成果之一
