Purchase behavior prediction model based on two-stage dynamic interest recognition

doi:10.11772/j.issn.1001-9081.2023081201

Journal of Computer Applications ›› 2024, Vol. 44 ›› Issue (8): 2365-2371.DOI: 10.11772/j.issn.1001-9081.2023081201

• Artificial intelligence • Previous Articles Next Articles

Purchase behavior prediction model based on two-stage dynamic interest recognition

Chunxue ZHANG, Liqing QIU(), Cheng’ai SUN, Caixia JING

Key Laboratory of Information Technology for Intelligent Mines in Shandong Province （Shandong University of Science and Technology），Qingdao Shandong 266590，China

Received:2023-09-05 Revised:2023-11-15 Accepted:2023-11-24 Online:2024-08-22 Published:2024-08-10
Contact: Liqing QIU
About author:ZHANG Chunxue， born in 1999， M. S. candidate. Her researchinterests include social network， user behavior prediction.
QIU Liqing ， born in 1978， Ph. D. ， associate professor. Herresearch interests include social networks.
SUN Cheng’ai ， born in 1964， M. S. ， professor. Her researchinterests include database， data mining.
JING Caixia， born in 1996， M. S. candidate. Her researchinterests include social network， user behavior prediction.
Supported by:
This work is partially supported by Natural Science Foundation ofShandong Province（ ZR2020MF044）.

基于两阶段动态兴趣识别的购买行为预测模型

张春雪, 仇丽青(), 孙承爱, 荆彩霞

山东省智慧矿山信息技术重点实验室（山东科技大学），山东青岛 266590

通讯作者: 仇丽青
作者简介:张春雪（1999—），女，山东济宁人，硕士研究生，主要研究方向：社交网络、用户行为预测
仇丽青（1978—），女，山东德州人，副教授，博士，主要研究方向：社交网络 qiuliqing2019@163.com
孙承爱（1964—），女，山东宁阳人，教授，硕士，主要研究方向：数据库、数据挖掘
荆彩霞（1996—），女，山东聊城人，硕士研究生，主要研究方向：社交网络、用户行为预测。
基金资助:
山东省自然科学基金资助项目(ZR2020MF044)

Abstract

Abstract:

Online purchase prediction aims to predict users’ purchase behaviors， which can generate considerable commercial value for shopping websites. To address the problem of traditional models’ inability to learn implicit interest preferences from users’ historical behaviors accurately， a two-stage dynamic interest recognition model for online purchase prediction was proposed to predict the probability of users purchasing products. Firstly， at the first stage of the model， a click frequency graph of user-product pairs was constructed， and Light-Graph Convolutional Network （LightGCN） was utilized to learn contextual features of the graph as the static interest’s representation of users. Then， at the second stage， Bidirectional Gated Recurrent Unit （Bi-GRU） with attention mechanism was applied to explore the transformation process of user preferences. Finally， aiming at the potential high-dimensional features， a purchase prediction model integrating dynamic interest and implicit features was built. The extensive experimental results on two real e-commerce datasets show that compared with Graph Convolutional Network （GCN） model， the proposed model has the accuracy improved by at least 0.3 percentage points， and the F1 score improved by at least 2.05 percentage points.

Key words: e-commerce, online purchase prediction, Light Graph Convolutional Neural Network (LightGCN), Bidirectional Gated Recurrent Unit (Bi-GRU), higher-order interest contextual feature

摘要：

在线购买预测旨在预测用户的购买行为，为购物网站带来可观的商业价值。针对传统模型学习用户历史行为中隐含的兴趣偏好不准确的问题，提出基于两阶段动态兴趣识别的购买行为预测模型，以预测用户购买商品的概率。首先，模型的第一阶段构建用户-商品的点击频率图，并利用轻量图卷积网络（LightGCN）学习图的上下文特征作为用户的静态兴趣表征；其次，第二阶段采用带有注意力机制的双向门控递归单元（Bi-GRU）探索用户偏好的转化过程；最后，针对潜在的高维特征，建立一个融合动态兴趣和隐含特征的购买预测模型。在2个真实电子商务数据集上的实验结果表明，所提模型与图卷积网络（GCN）模型相比，准确率至少提升0.3个百分点，F1分数至少提升了2.05个百分点。

关键词: 电子商务, 在线购买预测, 轻量图卷积神经网络, 双向门控递归单元, 高阶兴趣上下文特征

CLC Number:

TP391

Chunxue ZHANG, Liqing QIU, Cheng’ai SUN, Caixia JING. Purchase behavior prediction model based on two-stage dynamic interest recognition[J]. Journal of Computer Applications, 2024, 44(8): 2365-2371.

张春雪, 仇丽青, 孙承爱, 荆彩霞. 基于两阶段动态兴趣识别的购买行为预测模型[J]. 《计算机应用》唯一官方网站, 2024, 44(8): 2365-2371.

Figures/Tables 9

Tab. 1 Basic symbols

符号	描述	符号	描述
$G, υ, ϕ$	图，节点，边	$I u, i s$	用户的静态兴趣表示向量
$X h i s t t$	用户 $u$ 的交互序列	$R k u$	动态兴趣向量
$u, i$	用户，商品	$Q ν u$	兴趣的适应性表示向量
$W u, i τ$	边的权重	$H β$	高维隐式特征
$t$	交易记录的时间戳	$P (u, i)$	用户u购买商品i的概率

Tab. 1 Basic symbols

符号	描述	符号	描述
$G, υ, ϕ$	图，节点，边	$I u, i s$	用户的静态兴趣表示向量
$X h i s t t$	用户 $u$ 的交互序列	$R k u$	动态兴趣向量
$u, i$	用户，商品	$Q ν u$	兴趣的适应性表示向量
$W u, i τ$	边的权重	$H β$	高维隐式特征
$t$	交易记录的时间戳	$P (u, i)$	用户u购买商品i的概率

Fig. 1 Interaction between users and products

Fig. 2 Structure of purchase prediction model for two-stage dynamic interest recognition

Tab. 2 Description of datasets

属性	京东	淘宝
用户数	49 142	6 138
商品数	8 800	51 410
点击记录	54 756	632 791
正样本数^*	62 939	12 704
负样本数^**	484 621	618 599

Fig. 3 Influence of iterations on model performance on JD and Taobao datasets

Fig. 4 Influence of dropping rate on model performance on JD and Taobao dataset

Tab. 3 Ablation study results

模型	所有样本		正样本
模型	准确率	F1分数	准确率	F1分数
原始模型	93.68	82.38	61.70	68.64
模型（带Bi-GRU）	93.91	83.73	62.00	68.77
模型（带注意力机制）	93.94	84.21	62.63	70.26
模型（无权重矩阵）	94.07	84.32	63.18	72.09
模型（带有权重矩阵）	95.06	84.69	65.95	73.27

Tab. 4 Performance of proposed model on JD Dataset

模型	所有样本		正样本
模型	准确率	F1分数	准确率	F1分数
LINE-1st	92.67	72.54	54.74	56.99
LINE-2st	93.91	80.03	55.21	80.03
GCMC	93.85	80.25	56.69	80.25
NGCF	94.52	81.33	67.67	81.33
GCN	94.76	82.64	60.29	82.64
本文模型	95.06	84.69	65.95	84.69

Tab. 5 Performance of the proposed model on Taobao dataset

模型	所有样本		正样本
模型	准确率	F1分数	准确率	F1分数
LINE-1st	80.49	69.21	66.96	53.59
LINE-2st	82.72	71.45	68.26	55.68
GCMC	83.09	73.90	69.59	58.27
NGCF	84.27	75.67	71.50	59.12
GCN	85.03	76.08	73.76	64.15
本文模型	88.47	78.32	76.69	66.03

References 29

1	胡晓丽，张会兵，董俊超，等. 基于CNN-LSTM的用户购买行为预测模型［J］. 计算机应用与软件， 2020， 37（6）： 59-64.
	HU X L， ZHANG H B， DONG J C， et al. Prediction model of user buying behavior based on CNN-LSTM［J］. Computer Applications and Software， 2020， 37（6）： 59-64.
2	李美其，齐佳音. 基于购买行为及评论行为的用户购买预测研究［J］. 北京邮电大学学报（社会科学版）， 2016， 18（4）：18-25.
	LI M Q， QI J Y. Customer purchase prediction based on buying behavior and comment behavior［J］. Journal of Beijing University of Posts and Telecommunications （Social Sciences Edition）， 2016， 18（4）： 18-25.
3	张嵌嵌，何利力. 基于ResNet和DF融合的用户购买预测算法研究［J］. 软件工程与应用， 2022， 11（1）： 50-59.
	ZHANG Q Q， HE L L. Research on user purchase prediction algorithm based on the fusion of ResNet and DF［J］. Software Engineering and Applications， 2022， 11（1）： 50-59.
4	贾若雨，曾昂，朱敏，等. 面向在线交易日志的用户购买行为可视化分析［J］. 软件学报， 2017， 28（9）： 2450-2467.
	JIA R Y， ZENG A， ZHU M， et al. Visual analysis of consumer purchasing behavior for online transaction log［J］. Journal of Software， 2017， 28（9）： 2450-2467.
5	JIA R， LI R， YU M， et al. E-commerce purchase prediction approach by user behavior data［C］// Proceedings of 2017 International Conference on Computer， Information and Telecommunication Systems. Piscataway： IEEE， 2017： 1-5.
6	JIA R， LI R. Modeling user purchase preference based on implicit feedback［C］// Proceedings of 22nd IEEE International Conference on Computer Supported Cooperative Work in Design. Piscataway： IEEE， 2018： 832-836.
7	ZHANG Y， PANG L， SHI L， et al. Large scale purchase prediction with historical user actions on B2C online retail platform［EB/OL］. （2015-03-04）［2023-06-01］..
8	RESNICK P， IACOVOU N， SUCHAK M， et al. GroupLens： an open architecture for collaborative filtering of netnews［C］// Proceedings of the 1994 ACM Conference on Computer Supported Cooperative Work. New York： ACM， 1994： 175-186.
9	余小高，余小鹏. 基于隐式评分的推荐系统研究［J］. 计算机应用， 2009， 29（6）： 1585-1589.
	YU X G， YU X P. Research on recommendation system based on implicit rating［J］. Journal of Computer Applications， 2009， 29（6）： 1585-1589.
10	LIU X， LI J. Using support vector machine for online purchase predication［C］// Proceedings 2016 International Conference on Logistics， Informatics and Service Sciences. Piscataway： IEEE， 2016：1-6.
11	PARK C， KIM D， YANG M C， et al. Click-aware purchase prediction with push at the top［J］. Information Sciences， 2020， 521： 350-364.
12	ZHAO Y， YAO L， ZHANG Y. Purchase prediction using Tmall-specific features［J］. Concurrency and Computation： Practice and Experience， 2016， 28（14）： 3879-3894.
13	李振亮，李波. 基于矩阵分解的卷积神经网络改进方法［J］. 计算机应用， 2023， 43（3）： 685-691.
	LI Z L， LI B. Improved method of convolution neural network based on matrix decomposition［J］. Journal of Computer Applications， 2023， 43（3）： 685-691.
14	张继杰，杨艳，刘勇. 利用初始残差和解耦操作的自适应深层图卷积［J］. 计算机应用， 2022， 42（1）：9-15.
	ZHANG J J， YANG Y， LIU Y. Adaptive deep graph convolution using initial residual and decoupling operations［J］. Journal of Computer Applications， 2022， 42（1）： 9-15.
15	NIEPERT M， AHMED M， KUTZKOV K. Learning convolutional neural networks for graphs［C］// Proceedings of the 33rd International Conference on Machine Learning. New York： JMLR.org， 2016： 2014-2023.
16	KIPF T N， WELLING M. Semi-supervised classification with graph convolutional networks［EB/OL］. （2017-02-22）［2023-06-01］..
17	WU L， SUN P， HONG R， et al. SocialGCN： an efficient graph convolutional network based model for social recommendation［EB/OL］. （2019-07-11）［2023-06-01］..
18	VAN DEN BERG R， KIPFT N， WELLING F. Graph convolutional matrix completion［EB/OL］. （2017-10-25）［2023-06-01］..
19	WANG X， HE X， WANG M， et al. Neural graph collaborative filtering［C］// Proceedings of the 42nd International ACM SIGIR Conference on Research and Development in Information Retrieval. New York： ACM， 2019： 165-174.
20	HE X， DENG K， WANG X， et al. LightGCN： simplifying and powering graph convolution network for recommendation［C］// Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval. New York： ACM， 2020： 639-648.
21	LERCHE L， JANNACH D， LUDEWIG， M. On the value of reminders within e-commerce recommendations［C］// Proceedings of the 2016 Conference on User Modeling Adaptation and Personalization. New York： ACM， 2016： 27-35.
22	BAHDANAU D， CHO K， BENGIO Y. Neural machine translation by jointly learning to align and translate［EB/OL］. （2016-05-19）［2023-06-01］..
23	SONG K， YAO T， LING Q， et al. Boosting image sentiment analysis with visual attention［J］. Neurocomputing， 2018， 312： 218-228.
24	YAN X， HU S， MAO Y， et al. Deep multi-view learning methods： a review［J］. Neurocomputing， 2021， 448： 106-129.
25	ZHANG H， YAN J， ZHANG Y. CTR prediction models considering the dynamics of user interest［J］. IEEE Access， 2020， 8： 72847-72858.
26	MA R， HU X， ZHANG Q， et. al . Hot topic-aware retweet prediction with masked self-attentive model［C］// Proceedings of the 42nd International ACM SIGIR Conference on Research and Development in Information Retrieval. New York： ACM， 2019： 525-534.
27	RENDLE S， FREUDENTHALER C， GANTNER Z， et al. BPR： Bayesian personalized ranking from implicit feedback［C］// Proceedings of the 25th Conference on Uncertainty in Artificial Intelligence. Arlington， VA： AUAI Press， 2009： 452-461.
28	SHEN M， TANG C S， WU D， et al. JD.com： transaction-level data for the 2020 MSOM data driven research challenge［J］. Manufacturing and Service Operations Management， 2020， 26（1）：2-10.
29	TANG J， QU M， WANG M， et al. LINE： large-scale information network embedding［C］// Proceedings of the 24th International Conference on World Wide Web. Republic and Canton of Geneva： International World Wide Web Conferences Steering Committee， 2015： 1067-1077.

[1]	Liqing QIU, Xiaopan SU. Personalized multi-layer interest extraction click-through rate prediction model [J]. Journal of Computer Applications, 2024, 44(11): 3411-3418.
[2]	ZHANG Qi, JI Shujuan, FU Qiang, ZHANG Chunjin. Weighted reviewer graph based spammer group detection and characteristic analysis [J]. Journal of Computer Applications, 2019, 39(6): 1595-1600.
[3]	WANG Xuerong, WAN Nianhong. Dynamic prediction model on export sales based on controllable relevance big data of cross-border e-commerce [J]. Journal of Computer Applications, 2017, 37(4): 1038-1043.
[4]	LIU Dafu, SU Yang, XIE Hong'an, YANG Kai. Analysis of evolutionary game on client's evaluation strategy selection in e-commerce [J]. Journal of Computer Applications, 2016, 36(12): 3269-3273.
[5]	ZHOU Xiang, LI Shaobo, YANG Guanci. Entity recognition of clothing commodity attributes [J]. Journal of Computer Applications, 2015, 35(7): 1945-1949.
[6]	CAI Zhiwen, LIN Jianzong. Trust-aware collaborative filtering recommendation method for social E-commerce [J]. Journal of Computer Applications, 2015, 35(1): 167-171.
[7]	LIU Bin ZHANG Renjin. Promoting accuracy of trust bootstrapping from rating network [J]. Journal of Computer Applications, 2014, 34(8): 2442-2446.
[8]	FENG Yong LI Junping XU Hongyan DANG Xiaowan. Collaborative recommendation method improvement based on social network analysis [J]. Journal of Computer Applications, 2013, 33(03): 841-844.
[9]	LIU Xin LEI Wenqing. Fully anonymous multi-service subscription system without random oracles [J]. Journal of Computer Applications, 2013, 33(02): 417-429.
[10]	LU Zhi-gang ZHANG Xiao-xu. E-commerce reputation alliance profit distribution based on Shapley value method with cloud gravity center [J]. Journal of Computer Applications, 2012, 32(10): 2931-2934.
[11]	LIU Jia HUI Cheng-feng DU Xing-zhong CHEN Zhen-yu. E-commerce recommendation platform based on SaaS pattern [J]. Journal of Computer Applications, 2012, 32(09): 2679-2682.
[12]	XU Xiang-bin WANG Jia-qiang TU Huan MU Ming. Customer classification of E-commerce based on improved RFM model [J]. Journal of Computer Applications, 2012, 32(05): 1439-1442.
[13]	LIU Chao-fan ZHANG Yong-liang XIAO Gang. E-commerce security certification based on rolling fingerprint digital signature [J]. Journal of Computer Applications, 2012, 32(02): 475-479.
[14]	YANG Yang XIANG Yang XIONG Lei. Collaborative filtering and recommendation algorithm based on matrix factorization and user nearest neighbor model [J]. Journal of Computer Applications, 2012, 32(02): 395-398.
[15]	. Real-time recommendation method based on browsing preferences mining [J]. Journal of Computer Applications, 2011, 31(01): 89-92.

Purchase behavior prediction model based on two-stage dynamic interest recognition

基于两阶段动态兴趣识别的购买行为预测模型

RichHTML

PDF

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 9

References 29

Related Articles 15

Recommended Articles

Metrics