面向智能电网的多领域海量文本过滤框架研究

Multi-Domain Mass Text Filtering Framework for Smart Grid

高效安全的数据交互是实现智能电网信息化、互动化的重要保障,对智能电网互动数据进行识别与过滤是防止其泄露的关键技术,目前传统的集中式或分布式数据过滤方法已无法适应新的需求。文章提出面向智能电网的多领域海量文本过滤框架,重点描述了智能电网专业语料库的构建、智能电网多领域高维文本特征选择模型以及智能电网海量文本数据自适应分类等内容。文章提出的方法可以提高智能电网数据交换的安全性和准实时性,防止敏感业务数据的泄漏,对促进智能电网互动化业务应用安全稳定运行有重要意义。

Efficient and safe data exchange is an important guarantee of information development and interaction for smart grid, recognition and filtering of interactive data in smart grid is the key technical measure to prevent data leakage. To solve the problem that traditional centralized or distributed data filtering method can＇t meet the new requirement of smart grid, a multi-domain mass text filtering framework for smart grid is proposed in this paper. Smart grid professional corpus construction, multi-domain high dimensional text features selection model and mass text adaptive classification algorithm for smart grid are described in detail. The proposed framework can improve the safety and quasi real-time of smart grid, prevent sensitive business data leakage, which has important significance for improving the secure and stable operation of smart grid interactive business application.