From both a scientific and practical perspective, it is essential to evaluate how well NLG algorithms, models, and techniques work, using experiments that are rigorous and replicable. This chapter discusses basic evaluation concepts and then describes techniques for human and automatic (metric) evaluation. It also looks at evaluating real-world impact of NLG systems and commercial evaluation.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Evaluation

  • Ehud Reiter

摘要

From both a scientific and practical perspective, it is essential to evaluate how well NLG algorithms, models, and techniques work, using experiments that are rigorous and replicable. This chapter discusses basic evaluation concepts and then describes techniques for human and automatic (metric) evaluation. It also looks at evaluating real-world impact of NLG systems and commercial evaluation.