| 融合注意力机制与场感知因子分解机的电影推荐 |
| |
| 引用本文： | 黄德柒,苟 刚.融合注意力机制与场感知因子分解机的电影推荐[J].教育技术导刊,2020,19(6):19-22. |
| |
| 作者姓名： | 黄德柒 苟 刚 |
| |
| 作者单位： | 贵州大学 计算机科学与技术学院，贵州 贵阳 550025 |
| |
| 基金项目： | 国家自然科学基金项目（61562009） |
| |
| 摘 要： | 传统协同过滤算法仅利用评分信息进行推荐，而没有利用到更多用户特征与电影特征，推荐效果不佳。深度学习的普通应用，为特征提取打下了良好基础。通过爬取网站上的电影演员信息表，使用卷积神经网络对文本信息进行特征提取，采用结合注意力机制与场感知因子分解机的混合推荐方法，并使用用户—电影特征矩阵进行训练。在公开数据集 MovieLens 上进行实验测试，RMSE 达到 0.850，与 5 组推荐模型进行对比，RMSE 分别提18.0%、11.3%、7.60%、25.7%、6.80%。实验结果表明，该模型可以提高推荐效率。
|
| 关 键 词： | 推荐系统 注意力机制 场感知因子分解机 卷积神经网络 |
| 收稿时间： | 2019-08-13 |
Film Recommendation Based on Attention Mechanism and Field-aware Factorization Machine |
| |
| HUANG De-qi,GOU Gang.Film Recommendation Based on Attention Mechanism and Field-aware Factorization Machine[J].Introduction of Educational Technology,2020,19(6):19-22. |
| |
| Authors: | HUANG De-qi GOU Gang |
| |
| Institution: | College of Computer Science and Technology，Guizhou University，Guiyang 550025，China |
| |
| Abstract: | Traditional collaborative filtering algorithms only use rating information for recommendation，but do not use more user features and movie features，and fail to make good recommendation. With the general application of in-depth learning，it provides a good foundation for feature extraction. In this paper，we use convolutional neural network to extract feature from text information by crawling the corresponding information table of movie actors on the website. Combining the hybrid recommendation method of attention mechanism and field-aware factorization machine，we use user-film feature matrix for training. The RMSE reached 0.850 on the open data set MovieLens，which was compared with five recommended models，and achieved 18.0%，11.3%，7.60%，25.7% and 6.80% improvement. Experiments show that the proposed model can improve the efficiency of recommendation. |
| |
| Keywords: | recommendation system attention mechanism field-aware factoring machine convolutional neural network |
|
| 点击此处可从《教育技术导刊》浏览原始摘要信息 |
| 点击此处可从《教育技术导刊》下载免费的PDF全文 |
