耦合内容相似度信息的贝叶斯隐式反馈推荐

Bayesian Implicit Feedback Recommendation for Coupled Content Similarity Information

传统推荐算法如贝叶斯隐式反馈推荐算法都是围绕用户—态势需求度矩阵进行建模,只利用到用户和态势的交互行为信息,未利用到用户或态势的额外信息,而这些信息往往代表着用户对相关态势的潜在需求,充分利用这些额外信息,将有助于进一步提升刻画用户需求,提升用户建模的准确性。因此本章在贝叶斯隐式反馈推荐算法的基础上,融入用户和态势的相似度信息,提出一种耦合内容相似度信息的贝叶斯隐式反馈推荐算法。仿真结果表明,该算法能够充分挖掘用户与态势之间复杂的关系背后隐藏的用户潜在需求,提升战场态势推送的质量。

Traditional recommendation algorithms,such as Bayesian implicit feedback recommendation algorithms,are mod⁃eled around the user-situation demand matrix,which only uses the interaction behavior information of users and situations,and does not utilize additional information of users or situations,and such information often represents the potential needs of users for re⁃lated situations,making full use of this additional information will help to further enhance the user's needs and improve the accuracy of user modeling.Therefore,based on the Bayesian implicit feedback recommendation algorithm,this chapter integrates the similar⁃ity information of users and situation,and proposes a Bayesian implicit feedback recommendation algorithm for coupling content similarity information.The simulation results show that the algorithm can fully exploit the potential user hidden behind the complex relationship between users and situation,and improve the quality of battlefield situation push.