| 一种融合LDA主题模型与LSTM网络的个性化推荐算法 |
| |
| 引用本文： | 尧婉辰,孙怀远,谢润忠.一种融合LDA主题模型与LSTM网络的个性化推荐算法[J].教育技术导刊,2019,18(10):50-54. |
| |
| 作者姓名： | 尧婉辰 孙怀远 谢润忠 |
| |
| 作者单位： | 1.上海理工大学 医疗器械与食品学院，上海 200093；2.上海健康医学院 医疗器械学院，上海 201318；3.上海理工大学 光电信息与计算机工程学院，上海 200093 |
| |
| 摘 要： | 为改善传统个性化推荐算法精准度不高的问题，使用评论数据作为数据集，先对评论数据作文本预处理和特征提取，然后使用LDA主题模型对文本特征数据建模，得到主题词分布，将其作为标签，同时使用LSTM网络作文本分类，通过计算得到好评率。最后把用户需求和标签利用潜在语义标引计算相似度，根据相似度和好评率大小向用户推荐结果。实验结果表明，该方法能够向用户推荐符合其兴趣的个性化需求信息，且准确率高于96%，证明了该推荐算法的有效性。
|
| 关 键 词： | LDA主题模型 LSTM神经网络 个性化推荐算法 潜在语义标引 |
| 收稿时间： | 2018-12-24 |
Personalized Recommendation Algorithm Based on LDA Topic Model and LSTM Network |
| |
| YAO Wan-chen,SUN Huai-yuan,XIE Run-zhong.Personalized Recommendation Algorithm Based on LDA Topic Model and LSTM Network[J].Introduction of Educational Technology,2019,18(10):50-54. |
| |
| Authors: | YAO Wan-chen SUN Huai-yuan XIE Run-zhong |
| |
| Institution: | 1. School of Medical Instrument， Shanghai University of Science and Technology，Shanghai 200093，China；2. School of Medical Instrumention， Shanghai University of Medicine and Health Science，Shanghai 201318，China；3. School of Optical-Electrical and Computer Engineering， Shanghai University of Science and Technology，Shanghai 200093，China |
| |
| Abstract: | In order to improve the accuracy of traditional personalized recommendation algorithm， we used the comment data as the data set， and implemented the text preprocessing and feature extraction of the comment data. Then the LDA topic model was used to model the text feature data， and the subject word distribution was obtained， which was taken as a label. At the same time， LSTM network was used for text classification， and then the favorable comment rate was calculated. Finally， the similarity of user requirements and tags was calculated by potential semantic indexing， and the results were recommended to users according to the similarity and favorable comment rate. Experimental results show that this method can recommend personalized demand information in line with users' interests， and the accuracy is higher than 96%， which proves the effectiveness of this recommendation algorithm. |
| |
| Keywords: | LDA topic model LSTM neural network personalized recommendation algorithm latent semantic indexing |
|
| 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 |
