基于用户行为特征的深度混合推荐算法

doi:10.16088/j.issn.1001-6600.2023110603

摘要/Abstract

摘要： 现有大多数推荐方法学习的是每个特征的固定表示,然而用户行为偏好随上下文特征发生变化,特征在上下文中具有不同的重要性,因此,特征的固定表示造成模型给出的推荐结果不准确。为解决此问题,本文提出基于深度因式分解机并融合信息提取单元和交叉网络结构的混合推荐模型(deep and cross factorization machine information extraction unit,IEU-DeepCFM)。首先,自注意力机制和上下文信息提取器组成的信息提取单元模块对不同上下文中的每个特征学习上下文感知特征表示;然后,利用深度交叉因式分解机在提取用户低、高阶特征的同时来挖掘用户更多的显式交叉信息;最终实现对用户行为特征的点击率预测。在 MovieLens 电影数据集和Avazu广告点击率数据集上进行消融和对比实验,结果表明,本文所提出的模型在AUC和LogLoss 指标上均得到提高和改善,证明了该模型的合理性。

关键词: 深度学习, 上下文特征, 信息提取单元, 推荐算法, 自注意力机制

Abstract: The hybrid recommendation model,named IEU-DeepCFM (deep and cross factorization machine information extraction unit),is proposed in this paper,which is based on the deep factorization machine and integrates the information extraction unit and cross network structure. In the proposed model,a fixed representation of each feature is learned by most existing recommendation methods. However,it is recognized that user behavioral preferences change with contextual features,and features have different importance in different contexts. Therefore,inaccurate recommendation results may be caused by the fixed representation of features given by the model. To address this issue,the information extraction unit module is introduced,consisting of a self-attention mechanism and a contextual information extractor. This module learns context-aware feature representations for each feature in various contexts. Subsequently,a deep cross factorization machine is employed to mine low- and high-order features of the user. This enables users to receive more explicit cross-information,ultimately leading to click-through rate predictions based on user behavioral characteristics. The results of ablation and comparison experiments conducted on the MovieLens movie dataset and the Avazu advertising click-through rate dataset demonstrate the improvement in both AUC and LogLoss indicators achieved by the proposed model. This confirms the rationality of the model.

Key words: deep learning, contextual features, information extraction unit, recommendation algorithm, self-attention mechanism

中图分类号: TP391

杜帅文, 靳婷. 基于用户行为特征的深度混合推荐算法[J]. 广西师范大学学报（自然科学版）, 2024, 42(5): 91-100.

DU Shuaiwen, JIN Ting. A Deep Hybrid Recommendation Algorithm Based on User Behavior Characteristics[J]. Journal of Guangxi Normal University(Natural Science Edition), 2024, 42(5): 91-100.

参考文献

[1] 项亮. 推荐系统实践[M]. 北京: 人民邮电出版社, 2012: 5-10.
[2] 崔少国, 独潇, 杨泽田. 多注意力机制融合低高阶特征的神经推荐算法[J]. 计算机工程与应用, 2023, 59(8): 192-199. DOI: 10.3778/j.issn.1002-8331.2112-0570.
[3] 王瑞平, 贾真, 刘畅, 等. 基于DeepFM的深度兴趣因子分解机网络[J]. 计算机科学, 2021, 48(1): 226-232. DOI: 10.11896/jsjkx.191200098.
[4] 刘君良, 李晓光. 个性化推荐系统技术进展[J]. 计算机科学, 2020, 47(7): 47-55. DOI: 10.11896/jsjkx.200200114.
[5] 沈义峰, 金辰曦, 王瑶, 等. 融合时间上下文与特征级信息的推荐算法[J]. 计算机科学与探索, 2023, 17(2): 489-498. DOI: 10.3778/j.issn.1673-9418.2105008.
[6] 孔亚钰, 卢玉洁, 孙中天, 等. 面向强化当前兴趣的图神经网络推荐算法研究[J]. 广西师范大学学报(自然科学版), 2022, 40(3): 151-160. DOI: 10.16088/j.issn.1001-6600.2021071405.
[7] WANG F Y, WANG Y X, LI D S, et al. Enhancing CTR prediction with context-aware feature representation learning[C] //Proceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval. New York, NY: Association for Computing Machinery, 2022: 343-352. DOI: 10.1145/3477495.3531970.
[8] GUO H F, TANG R M, YE Y M, et al. DeepFM: a factorization-machine based neural network for CTR prediction[C] //Proceedings of the Twenty-Sixth International Joint Conference on Artificial Intelligence(IJCAI-17). Melbourne: IJCAI, 2017: 1725-1731. DOI: 10.24963/ijcai.2017/239.
[9] RENDLE S. Factorization machines[C] // 2010 IEEE International Conference on Data Mining. Los Alamitos, CA: IEEE Computer Society, 2010: 995-1000. DOI: 10.1109/ICDM.2010.127.
[10] WANG R X, FU B, FU G, et al. Deep & cross network for Ad click predictions[C] // Proceedings of the ADKDD'17. New York, NY: Association for Computing Machinery, 2017: 12. DOI: 10.1145/3124749.3124754.
[11] LIAN J X, ZHOU X H, ZHANG F Z, et al.xDeepFM: combining explicit and implicit feature interactions for recommender systems[C] // Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. New York, NY: Association for Computing Machinery, 2018: 1754-1763. DOI: 10.1145/3219819.3220023.
[12] SONG W P, SHI C C, XIAO Z P, et al. AutoInt: automatic feature interaction learning via self-attentive neural networks[C] // Proceedings of the 28th ACM International Conference on Information and Knowledge Management. New York, NY: Association for Computing Machinery, 2019: 1161-1170. DOI: 10.1145/3357384.3357925.
[13] ADOMAVICIUS G, TUZHILIN A. Context-Aware recommender systems[M] // RICCI F, ROKACH L, SHAPIRA B, et al. Recommender systems handbook. Boston: Springer, 2011: 217-253.
[14] 陈彬, 张荣梅, 张琦. DCFM: 基于深度学习的混合推荐模型[J]. 计算机工程与应用, 2021, 57(3): 150-155. DOI: 10.3778/j.issn.1002-8331.1910-0424.
[15] 刘亦欣, 王家伟, 李自力. 融合注意力与深度因子分解机的时间上下文推荐模型[J]. 计算机与现代化, 2021(11): 22-27. DOI: 10.3969/j.issn.1006-2475.2021.11.005.
[16] 郑浩. 融合上下文信息的深度推荐模型的研究[D]. 保定: 华北电力大学, 2021. DOI: 10.27139/d.cnki.ghbdu.2021.000395.
[17] 王利娥, 王艺汇, 李先贤. POI推荐中的多源数据融合和隐私保护方法[J]. 广西师范大学学报(自然科学版), 2023, 41(1): 87-101. DOI: 10.16088/j.issn.1001-6600.2022021104.
[18] CHEN Z, ZHONG F, CHEN Z, et al. DCAP: deep cross attentional product network for user response prediction[C] //Proceedings of the 30th ACM International Conference on Information and Knowledge Management. New York, NY: Association for Computing Machinery, 2021: 221-230. DOI: 10.1145/3459637.3482246.
[19] 唐铄丁. 基于深度兴趣网络的点击率预估模型的研究与实现[D]. 长春: 吉林大学, 2022. DOI: 10.27162/d.cnki.gjlin.2022.006196.
[20] ZHAO W X, MU S L, HOU Y P, et al.Recbole: towards a unified, comprehensive and efficient framework for recommendation algorithms[C] // Proceedings of the 30th ACM International Conference on Information & Knowledge Management. New York, NY: Association for Computing Machinery, 2021: 4653-4664. DOI: 10.1145/3459637.3482016.
[21] 邬彤, 于莲芝. 融入注意力网络的深度分解机推荐算法[J]. 电子科技, 2023, 36(1): 38-43, 50. DOI: 10.16180/j.cnki.issn1007-7820.2023.01.006.
[22] HE X N, CHUA T S. Neural factorization machines for sparse predictive analytics[C] // Proceedings of the 40th International ACM SIGIR conference on Research and Development in Information Retrieval. New York, NY: Association for Computing Machinery, 2017: 355-364. DOI: 10.1145/3077136.3080777.
[23] XIE R B, LING C, WANG Y L, et al. Deep feedback network for recommendation[C] // Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence(IJCAI-20). Yokohama: IJCAI, 2021: 2519-2525. DOI: 10.24963/ijcai.2020/349.
[24] XIAO J, YE H, HE X N, et al. Attentional factorization machines: learning the weight of feature interactions via attention networks[C] // Proceedings of the Twenty-Sixth International Joint Conference on Artificial Intelligence (IJCAI-17). Melbourne: IJCAI, 2017: 3119-3125. DOI: 10.24963/ijcai.2017/435.
[25] QU Y R, CAI H, REN K, et al. Product-Based neural networks for user response prediction[C] // 2016 IEEE 16th International Conference on Data Mining (ICDM). Los Alamitos, CA: IEEE Computer Society, 2016: 1149-1154. DOI: 10.1109/ICDM.2016.0151.
[26] 杨琳. 基于改进DeepFM模型的广告点击率预测研究[D]. 济南: 山东大学, 2022. DOI: 10.27272/d.cnki.gshdu.2022.003856.
[27] 周芙名. 基于深度学习和用户行为序列的推荐系统研究[D]. 长春: 吉林大学, 2022. DOI: 10.27162/d.cnki.gjlin.2022.006650.
[28] ZHANG W N, QIN J R, GUO W, et al. Deep learning for click-through rate estimation[C] // Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence (IJCAI-21). Montreal: IJCAI, 2021: 4695-4703. DOI: 10.24963/ijcai.2021/636.
[29] LI Z Y, CHENG W, CHEN Y, et al. Interpretable click-through rate prediction through hierarchical attention[C] //Proceedings of the 13th International Conference on Web Search and Data Mining. New York, NY: Association for Computing Machinery, 2020: 313-321. DOI: 10.1145/3336191.3371785.
[30] 黄立威, 江碧涛, 吕守业, 等. 基于深度学习的推荐系统研究综述[J]. 计算机学报, 2018, 41(7): 1619-1647. DOI: 10.11897/SP.J.1016.2018.01619.
[31] 吴正清, 曹晖, 刘宝锴. 基于注意力卷积神经网络的中文虚假评论检测[J]. 广西师范大学学报(自然科学版), 2023, 41(5): 26-36. DOI: 10.16088/j.issn.1001-6600.2023020502.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed