自动标引研究进展

Progress in Automatic Indexing Research

近五年来自动标引研究在关键词抽取、标引系统设计、自动分类标引、网络信息自动标引、数字图像标引、音频信息标引、视频信息标引、自动标引结果评价等方面取得很大进展,但尚存弱点与不足之处,还不能达到人工标引的效果。今后的研究将朝着探索更优越的语言分析技术、更高端的多媒体信息自动标引方法、高效的知识库智能自学习机制、多种标引方法或模型的互补的集成学习等方向发展。

Automatic indexing research, which includes keyword extraction, indexing systems design, automatic classification indexing, automatic indexing of network information, audio and video information, digital images, and evaluation of automatic indexing results, has made lots of progress during the last five years. But there is still deficiency in the research. Automatic indexing can not have the same effect of manual indexing. The future research may focus on the aspects such as superior technology of linguistic analysis, more advanced methods of automatic indexing of multimedia information, more effective self-learning mechanism of knowledge base and complementary ensemble learning of various indexing methods or models.