融合用户行为序列预测的混合推荐算法

doi:10.16180/j.cnki.issn1007-7820.2023.04.012

摘要/Abstract

摘要：

对于用户行为序列中隐藏用户兴趣的捕捉是近年来推荐算法研究的热门方向。传统的序列预测模型使用用户最后一次点击的商品为目标,建立用户行为和目标商品间的关联,并没有充分挖掘用户序列间的先后关系。文中在传统的DIN模型的基础上进行了改进,采用一段时间内的连续行为作为目标向量,使用transformer结构完成序列到序列的预测任务,进一步提取和利用了用户行为序列中的用户深度兴趣,并将其作为辅助特征结合DIN进行推荐。在亚马逊图书数据集和电子数据集上的实验结果表明,文中提出的基于DIN混合推荐模型比原DIN模型的AUC指标分别提升了约0.7%和1.9%。由此可知,基于用户行为序列预测的混合推荐可以在多特征推荐系统中起到一定的辅助作用。此外,文中还对用户序列长度对模型结果造成的影响进行了探究。

Abstract:

The capture of user interest hidden in the user behavior sequence is a hot research direction of recommendation algorithms in recent years. The traditional sequence prediction model uses the last product clicked by the user as the target, and establishes the association between user behavior and the target product, but does not fully dig out the sequence relationship between user sequences. This study improves on the traditional DIN model, uses continuous behavior over a period of time as the target vector, uses the transformer structure to complete the sequence-to-sequence prediction task, and further extracts and utilizes the user's deep interest in the user behavior sequence, and it is recommended in conjunction with DIN as an auxiliary feature. The experimental results on the Amazon book and the electronic data sets show that the DIN-based hybrid recommendation model proposed in this study increases the AUC index of the original DIN model by about 0.7% and 1.9%, respectively. It can be seen that the hybrid recommendation based on user behavior sequence prediction can play a certain auxiliary role in the multi-feature recommendation system. In addition, the influence of user sequence length on the model results is also explored.

Key words: recommendation system, click-through rate estimation, advertising calculation, CTR estimation, mixed recommendation, user sequence, user preference, attention mechanism

中图分类号:

TP391

孙红,鹿梅珂. 融合用户行为序列预测的混合推荐算法[J]. 电子科技, 2023, 36(4): 84-89.

SUN Hong,LU Meike. Hybrid Recommendation Algorithm Fused with User Behavior Sequence Prediction[J]. Electronic Science and Technology, 2023, 36(4): 84-89.

图/表 8

图1

图2

图3

表1

表2

图4

图5

图6

参考文献 18

[1]	颜金尧, 张海龙, 苏毓敏. 计算广告中的点击率和转化率预测研究[J]. 中国传媒大学学报(自然科学版), 2021, 28(2):54-60.
	Yan Jinyao, Zhang Hailong, Su Yumin. Recent research on the click-through rate and conversion rate prediction in computational advertising[J]. Journal of Communication University of China(Science and Technology), 2021, 28(2):54-60.
[2]	Rendle S. Factorization machines[C]. Sydney: Proceedings of the Tenth IEEE International Conference on Data Mining, 2010.
[3]	Guo H, Tang R, Ye Y, et al. DeepFM: A factorization-machine based neural network for CTR prediction[C]. Melbourne: Proceedings of the Twenty-sixth International Joint Conference on Artificial Intelligence, 2017.
[4]	Lian J, Zhou X, Zhang F, et al. xDeepFM: Combining explicit and implicit feature interactions for recommender systems[C]. London: Proceedings of the Twenty-fourth ACM SIGKDD International Conference, 2018.
[5]	郑萌. 基于改进注意力机制模型的智能英语翻译方法研究[J]. 电子科技, 2020, 33(11):84-87.
	Zheng Meng. Research on intelligent English translation method based on improved attention mechanism model[J]. Electronic Science and Technology, 2020, 33(11):84-87.
[6]	Zhou G, Zhu X, Song C, et al. Deep interest network for click-through rate prediction[C]. London: Proceedings of the Twenty-fourth ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, 2018.
[7]	Vaswani A, Shazeer N, Parmar N, et al. Attention is all you need[C]. Vancouver: Advances in Neural Information Processing Systems, 2017.
[8]	Zhou G, Mou N, Fan Y, et al. Deep interest evolution network for click-through rate prediction[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2019, 33(1):5941-5948. doi: 10.1609/aaai.v33i01.33015941
[9]	包晓安, 陈昀, 张瑞林, 等. 基于用户历史序列的点击率预估[J]. 软件导刊, 2021, 20(5):25-28.
	Bao Xiaoan, Chen Yun, Zhang Ruilin, et al. Click-through rate prediction based on user historical sequence[J]. Software Guide, 2021, 20(5):25-28.
[10]	曾彦程. 基于注意力机制的推荐算法研究[D]. 成都: 电子科技大学, 2021.
	Zeng Yancheng. Research on recommendation algorithm based on attention mechanism[D]. Chengdu: University of Electronic Science and Technology of China, 2021.
[11]	Sun F, Liu J, Wu J, et al. BERT4Rec: Sequential recommendation with bidirectional encoder representations from transformer[C]. Beijing: Proceedings of the Twenty-eighth ACM International Conference on Information and Knowledge Management, 2019.
[12]	王学斌. 一种基于深度偏好网络的点击率预估模型的设计与评价[D]. 北京: 北京大学, 2019.
	Wang Xuebin. Design and evaluation of a click-through rate prediction model based on deep preference network[D]. Beijing: Peking University, 2019.
[13]	Xiao Z, Yang L, Jiang W, et al. Deep multi-interest network for click-through rate prediction[C]. New York: Proceedings of the Twenty-ninth ACM International Conference on Information and Knowledge Management, 2020.
[14]	周菲, 徐洪珍. 基于改进Transformer的广告点击率预估模型[J]. 计算机应用研究, 2021, 38(8):2386-2389.
	Zhou Fei, Xu Hongzhen. Improved transformer based model for click-through rate prediction[J]. Application Research of Computers, 2021, 38(8):2386-2389.
[15]	李雪婷, 杨抒, 赛亚热·迪力夏提, 等. 融合内容与协同过滤的混合推荐算法应用研究[J]. 计算机技术与发展, 2021, 31(10):24-29.
	Li Xueting, Yang Shu, Saiyare Dilixiati, et al. Research on application of hybrid recommendation algorithm of content fusion and collaborative filtering[J]. Computer Technology and Development, 2021, 31(10):24-29.
[16]	Ma X, Zhao L, Huang G, et al. Entire space multi-task model: An effective approach for estimating post-click conversion rate[C]. Ann Arbor: Proceedings of the Forty-first International ACM SIGIR Conference, 2018.
[17]	Ni Y, Ou D, Liu S, et al. Perceive your users in depth: Learning universal user representations from multiple ecommerce tasks[C]. London: Proceedings of the Twenty-fourth ACM SIGKDD International Conference, 2018.
[18]	黄剑波, 陈方灵, 丁友东, 等. 基于情感分析的个性化电影推荐[J]. 计算机技术与发展, 2020, 30(9):132-136.
	Huang Jianbo, Chen Fangling, Ding Youdong, et al. Personalized movie recommendation based on sentiment Analysis[J]. Computer Technology and Development, 2020, 30(9):132-136.

模型参数	参数值
BatchSize	32
序列长度	15
多头注意力头数	2
Block	1
优化器	Adam

模型	AUC (books)	AUC (electronic)
BaseModel	0.708 32	0.677 24
DIN	0.748 45	0.709 13
DIEN	0.750 11	0.710 44
混合模型	0.753 67	0.722 39