一种基于神经网络模型的多检索词用户兴趣模型

A User Interest Model of Multiple Search Terms Based on Neural Network Model

针对职业教育资源库建设平台(职教云)目前对用户个性化推荐需求日益迫切的问题,提出了一种基于个性化神经网络的多检索词用户兴趣模型。通过将用户输入多检索词转换为词向量的形式,并将待推荐文档通过Doc2vec也转换为文档向量,二者通过个性化神经网络模型的相应特性,从而深层次挖掘出多检索词与用户兴趣之间的隐式联系,得到个性化推荐结果。实验结果表明,该方法从推荐效果上能够取得较好地效果。

Aiming at the urgent need of personalized recommendation for users in the construction platform of vocational education resource bank (vocational education cloud), a multi-search term user interest model based on personalized neural network is proposed. By transforming user input multi-search words into word vectors and recommending documents into document vectors through Doc2vec, the implicit relationship between multi-search words and user interests can be deeply excavated through the corresponding characteristics of personalized neural network model, and personalized recommendation results can be obtained. The experimental results show that the proposed method can achieve better recommendation results.