基于注意力机制的神经网络贝叶斯群组推荐算法

Neural Network Bayes Group Recommendation Algorithm Based on Attention Mechanism

目前基于注意力机制的神经网络推荐算法,如历史反馈信息的推荐系统作为深度学习推荐领域的重点课题,研究的是用户偏好学习同等影响,假设完整表达有效用户偏好特征,将注意力机制融入到深度学习推荐算法中,提出在历史反馈信息中强化注意力,基于注意力机制的推荐算法,输入网络中群组成员的个性化偏好,得到用户历史特征表示,利用多层神经网络对用户项目交互矩阵进行重塑,采用协同过滤的方法,得出用户反馈信息之间的依赖信息,突出用户个性化特征。这种基于深度学习和注意力机制的推荐算法,经过实验证明,对历史交互信息的耦合关系予以论证,得出深度学习推荐方法下注意力机制的推算效果。

At present,the neural network recommendation Algorithm based on attention mechanism,such as the history feedback information recommendation system,is a key topic in the field of deep learning recommendation,assuming that the feature of effective user preference is fully expressed,the attention mechanism is integrated into the deep learning recommendation algorithm,and a recommendation algorithm based on the attention mechanism is proposed to enhance the attention in the historical feedback information,by inputting the individual preferences of the group members in the network,we get the representation of the users'historical features,reconstruct the user item interaction matrix by using multi-layer neural networks,and obtain the dependent information among the users'feedback information by using the collaborative filtering method,give prominence to the user's personal characteristics.The Algorithm based on deep learning and attention mechanism is proved by experiments,and the coupling relationship of history mutual information is demonstrated,and the effect of attention mechanism is obtained.