Please wait a minute...
Advanced Search
数据分析与知识发现  2021, Vol. 5 Issue (3): 78-87     https://doi.org/10.11925/infotech.2096-3467.2020.0328
  研究论文 本期目录 | 过刊浏览 | 高级检索 |
融合近邻评论的GRU商品推荐模型*
冯勇1,刘洋1,徐红艳1,王嵘冰1(),张永刚2
1辽宁大学信息学院 沈阳 110036
2吉林大学符号计算与知识工程教育部重点实验室 长春 130012
Recommendation Model Incorporating Neighbor Reviews for GRU Products
Feng Yong1,Liu Yang1,Xu Hongyan1,Wang Rongbing1(),Zhang Yonggang2
1College of Information, Liaoning University, Shenyang 110036, China
2Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education,Jilin University, Changchun 130012, China
全文: PDF (918 KB)   HTML ( 6
输出: BibTeX | EndNote (RIS)      
摘要 

【目的】 改进矩阵分解算法,融合近邻评论缓解评论稀疏,提高推荐准确性。【方法】 应用多层感知机改进矩阵分解算法以得到用户和商品的深层非线性特征;处理评论集并深度融合近邻特征与用户特征,表征更符合用户偏好的特征;基于所得特征预测评分进行推荐。【结果】 在亚马逊数据集上与其他深度模型进行对比实验,在准确率、召回率、归一化折损累计增益指标上最多可提升8.3%、22.8%、14.9%。【局限】 未考虑用户给出评论的时间因素,并且忽略可能存在虚假评论的情况。【结论】 改进的矩阵分解算法和融合近邻评论的方法可以有效提升推荐的准确性。

服务
把本文推荐给朋友
加入引用管理器
E-mail Alert
RSS
作者相关文章
冯勇
刘洋
徐红艳
王嵘冰
张永刚
关键词 推荐模型评分评论稀疏性深度学习    
Abstract

[Objective] This paper improves the matrix factorization algorithm with neighboring user’s comments, aiming to address the sparse comments issue and improve recommendation accuracy. [Methods] First, we used the Multi-layer Perceptron to improve the matrix decomposition algorithm and obtain the deep nonlinear features of users and commodities. Then, we processed the reviews and integrated the characteristics of users and their neighbors. Third, we identified users’ features in line with their preferences. Finally, we made recommendations based on the obtained prediction scores of the features. [Results] We compared the performance of our new algorithm with other models on the Amazon dataset. The accuracy, recall, and normalized cumulative loss gain of the proposed model increased by up to 8.3%, 22.8%, and 14.9%, respectively. [Limitations] We neither included the time factor of the user’s comments, nor excluded the fake comments. [Conclusions] Our new algorithm could effectively improve the recommendation results.

Key wordsRecommendation Model    Rating    Reviews    Sparsity    Deep Learning
收稿日期: 2020-10-22      出版日期: 2020-11-24
ZTFLH:  TP302  
  G202  
基金资助:*教育部重点实验室资助项目(93K172018K01);辽宁省社会科学规划基金项目(L18AGL007);赛尔网络下一代互联网技术创新项目(NGII20190301)
通讯作者: 王嵘冰     E-mail: wrb@lnu.edu.cn
引用本文:   
冯勇,刘洋,徐红艳,王嵘冰,张永刚. 融合近邻评论的GRU商品推荐模型*[J]. 数据分析与知识发现, 2021, 5(3): 78-87.
Feng Yong,Liu Yang,Xu Hongyan,Wang Rongbing,Zhang Yonggang. Recommendation Model Incorporating Neighbor Reviews for GRU Products. Data Analysis and Knowledge Discovery, 2021, 5(3): 78-87.
链接本文:  
https://manu44.magtech.com.cn/Jwk_infotech_wk3/CN/10.11925/infotech.2096-3467.2020.0328      或      https://manu44.magtech.com.cn/Jwk_infotech_wk3/CN/Y2021/V5/I3/78
Fig.1  GRU-NR模型框架
Fig.2  深度融合用户特征与近邻特征
数据集 用户数 商品数 样本数 稀疏度 用户发表平均评论数 商品收到平均评论数
Toys_and_Games 19 412 11 924 167 597 99.92% 8.63 14.05
Digital_Music 5 541 3 568 64 706 99.67% 11.67 18.14
Automotive 2 928 1 835 20 473 99.62% 7.23 20.35
Table 1  数据集基本信息
数据集 LFM DTMF NARRE HRDR DeepCLFM GRU-NR
Toys_and_Games 1.246 2 1.041 6 0.982 6 0.934 0 0.926 1 0.901 0
Digital_Music 1.095 9 0.963 8 0.912 5 0.897 3 0.871 3 0.865 3
Automotive 1.168 8 1.008 0 0.943 4 0.911 6 0.912 4 0.894 2
Table 2  评分预测准确度对比
Fig.3  各模型生成推荐列表的召回率
Fig.4  各模型生成推荐列表的NDCG值
Fig.5  近邻评论适用性分析
方法 方法描述 相似度(偏好) 运算速度排序 是否选择其他商品的评论 差异性 推荐准确性(RMSE)
方法1 选出与当前商品评分一致的用户,直接将其评论作为近邻评论 0.372 1 0.951 9
方法2 选出与当前商品评论语义相似度最高的评论作为近邻评论 0.796 3 0.933 7
方法3 本文方法 0.715 2 0.894 2
Table 3  近邻评论选择方法对比
[1] 侯银秀, 李伟卿, 王伟军, 等. 基于用户偏好与商品属性情感匹配的图书个性化推荐研究[J]. 数据分析与知识发现, 2017,1(8):9-17.
[1] ( Hou Yinxiu, Li Weiqing, Wang Weijun, et al. Personalized Book Recommendation Based on User Preferences and Commodity Features[J]. Data Analysis and Knowledge Discovery, 2017,1(8):9-17.)
[2] Zheng L, Noroozi V, Yu P S. Joint Deep Modeling of Users and Items Using Reviews for Recommendation[C]// Proceedings of the 10th ACM International Conference on Web Search and Data Mining. ACM, 2017: 425-434.
[3] LeCun Y, Bottou L, Bengio Y, et al. Gradient-Based Learning Applied to Document Recognition[J]. Proceedings of the IEEE, 1998,86(11):2278-2324.
[4] Xing S N, Liu F A, Wang Q Q, et al. A Hierarchical Attention Model for Rating Prediction by Leveraging User and Product Reviews[J]. Neurocomputing, 2019,332(3):417-427.
[5] 邢长征, 赵宏宝, 张全贵, 等. 融合评论文本层级注意力和外积的推荐方法[J]. 计算机科学与探索, 2020,14(6):947-957.
[5] ( Xing Changzheng, Zhao Hongbao, Zhang Quangui, et al. Review Text Hierarchical Attention and Outer Product for Recommendation Method[J]. Journal of Frontiers of Computer Science and Technology, 2020,14(6):947-957.)
[6] Koren Y, Bell R, Volinsky C. Matrix Factorization Techniques for Recommender Systems[J]. Computer, 2009,42(8):30-37.
[7] Rosenblatt F. The Perceptron: A Perceiving and Recognizing Automaton[R]. Technical Report 85-460-1 of Project PARA. New York,USA: Cornell Aeronautical Laboratory, 1957.
[8] Chung J, Gulcehre C, Cho K H, et al. Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling[OL]. arXiv Preprint, arXiv: 1412. 3555.
[9] Mnih A, Salakhutdinov R R. Probabilistic Matrix Factorization[C]// Proceedings of the Advances in Neural Information Processing Systems, Vancouver, Canada. Cambridge, MA: MIT Press, 2008: 1257-1264.
[10] 宋瑞雪, 李国勇. 基于改进的矩阵分解模型在推荐系统中的应用[J]. 计算机应用, 2019,39(S1):93-95.
[10] ( Song Ruixue, Li Guoyong. Application of Improved Matrix Factorization Model in Recommendation System[J]. Journal of Computer Applications, 2019,39(S1):93-95.)
[11] 温彦, 马立健, 曾庆田, 等. 基于地理信息偏好修正和社交关系偏好隐式分析的POI推荐[J]. 数据分析与知识发现, 2019,3(8):30-40.
[11] ( Wen Yan, Ma Lijian, Zeng Qingtian, et al. POI Recommendation Based on Geographic and Social Relationship Preferences[J]. Data Analysis and Knowledge Discovery, 2019,3(8):30-40.)
[12] 李琳, 刘锦行, 孟祥福, 等. 融合评分矩阵与评论文本的商品推荐模型[J]. 计算机学报, 2018,41(7):1559-1573.
[12] ( Li Lin, Liu Jinxing, Meng Xiangfu, et al. Recommendation Models by Exploiting Rating Matrix and Review Text[J]. Chinese Journal of Computers, 2018,41(7):1559-1573.)
[13] Chen C, Zhang M, Liu Y Q. Neural Attentional Rating Regression with Review-level Explanations[C]// Proceedings of the 2018 World Wide Web Conference. 2018: 1583-1592.
[14] Wu H, Zhang Z X, Yue K, et al. Dual-Regularized Matrix Factorization with Deep Neural Networks for Recommender Systems[J]. Knowledge-Based Systems, 2018,145:46-58.
[15] Pennington J, Socher R, Manning C. GloVe: Global Vectors Forword Representation[C]// Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing. 2014: 1532-1543.
[16] Devlin J, Chang M W, Lee K, et al. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding[OL]. arXiv Preprint, arXiv: 1810. 04805.
[17] 冯兴杰, 曾云泽. 基于评分矩阵与评论文本的深度推荐模型[J]. 计算机学报, 2020,43(5):884-900.
[17] ( Feng Xingjie, Zeng Yunze. Joint Deep Modeling of Rating Matrix and Reviews for Recommendation[J]. Chinese Journal of Computers, 2020,43(5):884-900.)
[18] Feng X J, Zeng Y Z. Neural Collaborative Embedding from Reviews for Recommendation[J]. IEEE Access, 2019,7:103263-103274.
[19] Liu H T, Wang Y A, Peng Q Y, et al. Hybrid Neural Recommendation with Joint Deep Representation Learning of Ratings and Reviews[J]. Neurocomputing, 2020,374:77-85.
[20] 冯兴杰, 曾云泽, 崔桂颖. 基于近邻用户评论的推荐辅助网络[J]. 计算机应用研究, 2020,37(10):2956-2960.
[20] ( Feng Xingjie, Zeng Yunze, Cui Guiying. Recommendation Supplemental Network of Neighbor Reviews[J]. Application Research of Computers, 2020,37(10):2956-2960.)
[21] Gerard S, Christopher B. Term-Weighting Approaches in Automatic Text Retrieval[J]. Information Processing and Management, 1988,24(5):513-523.
[22] Rendle S. Factorization Machines[C]// Proceedings of 2010 IEEE International Conference on Data Mining. 2010: 995-1000.
[1] 周泽聿,王昊,赵梓博,李跃艳,张小琴. 融合关联信息的GCN文本分类模型构建及其应用研究*[J]. 数据分析与知识发现, 2021, 5(9): 31-41.
[2] 韩辉, 刘秀文. 海事适任评估中主观题自动评分技术研究*[J]. 数据分析与知识发现, 2021, 5(8): 113-121.
[3] 赵丹宁,牟冬梅,白森. 基于深度学习的科技文献摘要结构要素自动抽取方法研究*[J]. 数据分析与知识发现, 2021, 5(7): 70-80.
[4] 徐月梅, 王子厚, 吴子歆. 一种基于CNN-BiLSTM多特征融合的股票走势预测模型*[J]. 数据分析与知识发现, 2021, 5(7): 126-138.
[5] 黄名选,蒋曹清,卢守东. 基于词嵌入与扩展词交集的查询扩展*[J]. 数据分析与知识发现, 2021, 5(6): 115-125.
[6] 钟佳娃,刘巍,王思丽,杨恒. 文本情感分析方法及应用综述*[J]. 数据分析与知识发现, 2021, 5(6): 1-13.
[7] 马莹雪,甘明鑫,肖克峻. 融合标签和内容信息的矩阵分解推荐方法*[J]. 数据分析与知识发现, 2021, 5(5): 71-82.
[8] 张国标,李洁. 融合多模态内容语义一致性的社交媒体虚假新闻检测*[J]. 数据分析与知识发现, 2021, 5(5): 21-29.
[9] 李菲菲,吴璠,王中卿. 基于生成式对抗网络和评论专业类型的情感分类研究 *[J]. 数据分析与知识发现, 2021, 5(4): 72-79.
[10] 胡昊天,吉晋锋,王东波,邓三鸿. 基于深度学习的食品安全事件实体一体化呈现平台构建*[J]. 数据分析与知识发现, 2021, 5(3): 12-24.
[11] 张琪,江川,纪有书,冯敏萱,李斌,许超,刘浏. 面向多领域先秦典籍的分词词性一体化自动标注模型构建*[J]. 数据分析与知识发现, 2021, 5(3): 2-11.
[12] 吕学强,罗艺雄,李家全,游新冬. 中文专利侵权检测研究综述*[J]. 数据分析与知识发现, 2021, 5(3): 60-68.
[13] 常城扬,王晓东,张胜磊. 基于深度学习方法对特定群体推特的动态政治情感极性分析*[J]. 数据分析与知识发现, 2021, 5(3): 121-131.
[14] 成彬,施水才,都云程,肖诗斌. 基于融合词性的BiLSTM-CRF的期刊关键词抽取方法[J]. 数据分析与知识发现, 2021, 5(3): 101-108.
[15] 李丹阳, 甘明鑫. 基于多源信息融合的音乐推荐方法 *[J]. 数据分析与知识发现, 2021, 5(2): 94-105.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
版权所有 © 2015 《数据分析与知识发现》编辑部
地址:北京市海淀区中关村北四环西路33号 邮编:100190
电话/传真:(010)82626611-6626,82624938
E-mail:jishu@mail.las.ac.cn