基于内容相似度的雷达情报筛选技术

Radar information screening technology based on content similarity

为了实现雷达情报信息的按需推送、解决情报用户信息过载的问题,提出了一种利用内容相似度的个性化推荐技术筛选用户感兴趣情报信息的方法.通过专家评价建立基于情报特征向量的用户兴趣模型,利用余弦相似度和欧几里得相似度算法获得其内容相似度,根据相似度形成用户的雷达情报推荐.仿真分析比较了两种相似度算法对情报内容的处理性能,结果表明两种算法均可实现情报推荐,而欧几里得相似度算法具有更好的平均绝对偏差性能.

In order to realize the on-demand push of radar intelligence information and solve the problem as information overload of intelligence users, this paper proposes a method to screen information of interest to the users by utilizing the content similarity individuation recommendation technique. Through the expert evaluation,the paper establishes the user interest model based on the intelligence feature vector. The similarity degree is obtained by using the cosine similarity degree and the Euclidean similarity algorithm, and the user's radar information recommendation is formed according to the similarity degree. Finally the paper does the simulation analyses to compare the two similarity algorithms in the aspect of intelligence-processing performance. Results show that both algorithms can realize the recommendation of radar intelligence and that the Euclidean similarity algorithm has better performance on mean absolute error.