针对Internet上的信息过载问题,提出了一种基于内容分析的信息推荐方法。该方法使用神经网络作为知识表示和推理机制来建立用户兴趣模型,然后以用户模型为基础来预测用户对信息的偏好程度,并据此做出信息推荐。提出的方法通过一个仿真试验进行验证。
This paper proposes a new content-based information recommendation method to address the information overload problem on Internet.First,the proposed method constructs a user profile to characterize user interests,and then predicts the unknown preferences of the user has on new items based on constructed user profile.The new items with the highest preference prediction values are recommended to the user.The whole recommendation process is supported by using neural networks as knowledge representation and inference mechanism.The proposed method is validated by a simulation experiment.