摘要
为了提高情报分发的效率,解决雷达组网上信息过载的问题,提出了一种基于朴素贝叶斯分类算法的雷达情报按需分发技术。利用层次向量空间构建用户兴趣空间,对情报用户的历史情报和定制信息,通过朴素贝叶斯分类算法挖掘用户兴趣,建立用户兴趣模型;通过实时情报与用户兴趣模型的匹配,将情报用户感兴趣的情报推送给用户,过滤其不感兴趣的情报,从而实现雷达情报的按需分发。仿真实验将该技术与基于TF-IDF分类算法的情报分发技术做了准确率与覆盖率的对比实验,结果表明,该方法的准确率优于利用TF-IDF分类算法的情报分发技术,能够较好地实现雷达情报的按需分发。
For improving the efficiency of intelligence distribution, and solving the problem of the information overload on the radar intelligence network, the technology of filtering the users＇ interested intelligence based on the personalized recommending techniques is proposed. The historical information and the registered data are utilized for Naive Bayes classifier to mine the users＇interests, which presents by the hierarchical vector space model. With the discovered user profile, the most matched intelligence is recommended to the users. And the results show that this algorithm can realize radar data distribution for the personalized demands, and have a better performance than the algorithm based on TF-IDF classifier.
出处
《现代雷达》
CSCD
北大核心
2014年第7期46-50,53,共6页
Modern Radar
关键词
情报按需分发
层次向量空间模型
用户兴趣挖掘
朴素贝叶斯
radar data distribution on demands
hierarchical vector space model
user interest mining
Naive Bayes
