摘要
报告是企业日常运营中具有重复性与复杂性的一项工作,本文系统梳理了报告自动生成的实现方法、研究重点与发展方向。在总结基于数据合并、模板等基本报告生成方法的基础上,首先介绍了序列生成、知识指导等智能化报告生成的改进模型;然后进一步分析了基于模板与智能化的混合模型报告生成方法;最后对报告生成的评价标准进行了对比。研究表明:基于模板与智能化的报告生成方法具有程式化、精度高、适应性强等特点,是今后研究的主流方向。
Reporting is a repetitive and complex task in the daily operation of enterprises. In this paper, we systematically sort out the implementation methods, research priorities and development directions of automatic report generation. This paper summarizes the basic report generation methods based on data merging and templates and introduces the improved models of intelligent report generation such as sequence generation and knowledge guidance, then further analyzes the mixed model report generation methods based on templates and intelligence;finally, the evaluation criteria of report generation are compared. We come to the conclusion that report generation methods based on templates and intelligence has the characteristics of programmatic, high accuracy, and adaptability. It will be the mainstream direction of future research.
出处
《计算机科学与应用》
2022年第10期2307-2317,共11页
Computer Science and Application
