Please wait a minute...
浙江大学学报(工学版)
Journal of ZheJiang University (Engineering Science)  2022, Vol. 56 Issue (2): 254-262    DOI: 10.3785/j.issn.1008-973X.2022.02.005
    
Uncertain behavior sequence prediction method based on intent identification
Fei HE1,2(),Cang-hong JIN1,Ming-hui WU1,*()
1. School of Computer and Computing Science, Zhejiang University City College, Hangzhou 310015, China
2. College of Computer Science and Technology, Zhejiang University, Hangzhou 310027, China
Download: HTML     PDF(1039KB) HTML
Export: BibTeX | EndNote (RIS)      

Abstract  

An graph based intent identification embedding (G2IE) method was proposed, in order to solve the problems of behavior uncertainty and data sparsity faced by collaborative recommendation and sequence representation methods in user behavior prediction. In G2IE method, firstly the theory of planned behavior (TPB) is used to mine the controlled behavior patterns in the user behavior sequence, then the transfer intention intensity of the uncertain behavior list between adjacent controlled behaviors is calculated based on information entropy, and finally the behavior relationship is strengthened by integrating the behavior transfer intention to make up for the lack of behavior intention. In G2IE method, the uncertainty of behavior is identified and it is measured with a model, in order to solve the problem of behavior randomness. The problem of data sparsity can be alleviated to some extent by discovering more behavior relationships through the fusion of transfer intention. G2IE method has more accurate and rich expression ability compared with other methods that use behavior direct relation. Experimental results on three public user behavior datasets demonstrate the effectiveness of the proposed method.

Key words： behavior pattern mining      uncertainty relationship      intent identification      graph embedding      behavior sequence prediction     
Received: 10 October 2021      Published: 03 March 2022
CLC:  TP 391.4  
Corresponding Authors: Ming-hui WU     E-mail: fei.he@zju.edu.cn;mhwu@zucc.edu.cn
Service
E-mail this article
Add to my bookshelf
Add to citation manager
E-mail Alert
RSS
Articles by authors
Fei HE
Cang-hong JIN
Ming-hui WU
Cite this article:

Fei HE,Cang-hong JIN,Ming-hui WU. Uncertain behavior sequence prediction method based on intent identification. Journal of ZheJiang University (Engineering Science), 2022, 56(2): 254-262.

URL:

https://www.zjujournals.com/eng/10.3785/j.issn.1008-973X.2022.02.005     OR     https://www.zjujournals.com/eng/Y2022/V56/I2/254


基于意图识别的不确定性行为序列预测方法

针对协同推荐和序列表征方法在预测用户行为任务上面临的行为不确定性和数据稀疏问题，提出基于意图识别的不确定性行为序列预测(G2IE)方法. G2IE方法根据计划行为理论(TPB)，对用户行为序列中受控行为模式进行挖掘；基于信息熵计算相邻受控行为之间的不确定性行为列表的行为转移意图强度；融合行为转移意图增强行为关系，弥补行为意图缺失. G2IE方法挖掘行为的不确定性关系，并用模型进行量化，用于解决行为不确定性难点；通过融合转移意图方法能够发现更多的行为关系，也在一定程度上缓解数据稀疏的问题. 较其他使用行为直接关系的方法，G2IE方法有更准确丰富的表示能力. 在3个公开行为数据集上进行对比实验，结果表明，本研究方法在综合指标F1值上均为最优，证明了所提方法的有效性.


关键词： 行为模式挖掘,  不确定性关系,  意图识别,  图嵌入,  行为序列预测 
Fig.1 Overall architecture of G2IE method
Fig.2 Transfer intention relations mining of controlled behavior
Fig.3 Example of intent identification behavior graph
数据集 $|U|$ $|B|$ ${\rm{len} }_{ {\text{a} } }$ $n\text{′}$ ${\rm{SI} }/\rm{\text{%} }$
ML 6040 3377 165.47 999416 95.10
RecSys 45520 4519 6.60 300105 99.85
Beauty 40037 13951 6.52 261205 99.95
Tab.1 Dataset statistics
数据集 ${n_{{\text{ori}}}}$ ${n_{ { {\rm{int} }} } }$ $\Delta n /{\text{% }}$
ML 375039 491539 31.06
RecSys 92901 103984 11.93
Beauty 133135 143 433 7.74
Tab.2 Difference of relation num between intent identification behavior graph and original behavior graph
数据集 方法 prec@5 recall@5 F1@5 prec@10 recall@10 F1@10
ML Random 0.00993 0.00146 0.00255 0.01061 0.00313 0.00483
MostPopular 0.11129 0.02134 0.03581 0.10109 0.03754 0.05475
ItemKNN 0.09152 0.02785 0.04271 0.08851 0.05254 0.06594
BPRMF 0.12543 0.03066 0.04928 0.11334 0.05379 0.07296
GRU4Rec 0.29106 0.02958 0.05370 0.26023 0.05290 0.08793
Caser 0.11815 0.03201 0.05037 0.12210 0.06371 0.08373
BGE 0.29099 0.02958 0.05369 0.25954 0.05276 0.08769
G2IE? 0.28758 0.02923 0.05307 0.25878 0.05260 0.08743
G2IE 0.29176 0.02965 0.05383 0.26209 0.05328 0.08855
RecSys Random 0.00011 0.00045 0.00018 0.00013 0.00098 0.00023
MostPopular 0.00308 0.01165 0.00487 0.00298 0.02280 0.00527
ItemKNN 0.02421 0.09515 0.03860 0.02495 0.19414 0.04422
BPRMF 0.00323 0.01256 0.00514 0.00327 0.02499 0.00578
GRU4Rec 0.02086 0.06909 0.03205 0.01666 0.11033 0.02895
Caser 0.03139 0.10862 0.04871 0.02804 0.19390 0.04899
BGE 0.03935 0.13032 0.06045 0.02959 0.19600 0.05142
G2IE? 0.03770 0.12486 0.05791 0.02877 0.19056 0.04999
G2IE 0.04410 0.14605 0.06774 0.03252 0.21540 0.05651
Beauty Random 0.00006 0.00014 0.00008 0.00005 0.00030 0.00009
MostPopular 0.00207 0.00520 0.00296 0.00182 0.00919 0.00304
ItemKNN 0.00182 0.00422 0.00254 0.00202 0.00906 0.00330
BPRMF 0.00205 0.00527 0.00295 0.00174 0.00875 0.00290
GRU4Rec 0.00338 0.00837 0.00482 0.00312 0.01545 0.00519
Caser 0.00440 0.00930 0.00597 0.00380 0.01610 0.00615
BGE 0.00606 0.01501 0.00864 0.00519 0.02567 0.00863
G2IE? 0.00618 0.01529 0.00880 0.00543 0.02686 0.00903
G2IE 0.00683 0.01690 0.00973 0.00598 0.02959 0.00995
Tab.3 Comparison of method performance on three datasets
Fig.4 Impact of behavior embedding dimensions on F1@10 indicator
[1]   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
FENG Xing-jie, ZENG Yun-ze Joint deep modeling of rating matrix and reviews for recommendation[J]. Chinese Journal of Computers, 2020, 43 (5): 884- 900
doi: 10.11897/SP.J.1016.2020.00884
[2]   陈彦敏, 王皓, 马建辉, 等 基于层级注意力机制的互联网用户信用评估框架[J]. 计算机研究与发展, 2020, 57 (8): 1755- 1768
CHEN Yan-min, WANG Hao, MA Jian-hui, et al A hierarchical attention mechanism framework for internet credit evaluation[J]. Journal of Computer Research and Development, 2020, 57 (8): 1755- 1768
doi: 10.7544/issn1000-1239.2020.20200217
[3]   GU Y, YANG X, PENG M, et al Robust weighted SVD-type latent factor models for rating prediction[J]. Expert Systems with Applications, 2020, 141: 112885
doi: 10.1016/j.eswa.2019.112885
[4]   EKSTRAND M D, RIEDL J T, KONSTAN J A. Collaborative filtering recommender systems [M]. Boston: Now Publishers Inc, 2011.
[5]   黄璐, 林川杰, 何军, 等 融合主题模型和协同过滤的多样化移动应用推荐[J]. 软件学报, 2017, 28 (3): 708- 720
HUANG Lu, LIN Chuan-jie, HE Jun, et al Diversified mobile app recommendation combining topic model and collaborative filtering[J]. Journal of Software, 2017, 28 (3): 708- 720
[6]   LIU D R, LAI C H, LEE W J A hybrid of sequential rules and collaborative filtering for product recommendation[J]. Information Sciences, 2009, 179 (20): 3505- 3519
doi: 10.1016/j.ins.2009.06.004
[7]   JARBOUI F, GRUSON-DANIEL C, DURMUS A, et al. Markov decision process for MOOC users behavioral inference [C]// European MOOCs Stakeholders Summit. Naples: Springer, 2019: 70-80.
[8]   HIDASI B, KARATZOGLOU A, BALTRUNAS L, et al. Session-based recommendations with recurrent neural networks[EB/OL]. [2021-10-10]. https://arxiv.org/abs/1511.06939.
[9]   CHO K, VAN MERRIËNB B, GULCEHRE C, et al. Learning phrase representations using RNN encoder-decoder for statistical machine translation [EB/OL]. [2021-10-10]. https://arxiv.org/abs/1406.1078.
[10]   YU Z, LIAN J, MAHMOODY A, et al. Adaptive user modeling with long and short-term preferences for personalized recommendation [C]// Proceedings of the 28th International Joint Conference on Artificial Intelligence. Macao: AAAI Press, 2019: 4213-4219.
[11]   SHEIL H, RANA O. Classifying and recommending using gradient boosted machines and vector space models [C]// UK Workshop on Computational Intelligence. Nottingham: Springer, 2017: 214-221.
[12]   TANG J, WANG K. Personalized top-n sequential recommendation via convolutional sequence embedding [C]// Proceedings of the 11th ACM International Conference on Web Search and Data Mining. [S.l.] : ACM, 2018: 565-573.
[13]   刘浩翰, 吕鑫, 李建伏 考虑用户意图和时间间隔的会话型深度学习推荐系统[J]. 计算机应用与软件, 2021, 38 (3): 190- 195
LIU Hao-han, LV Xin, LI Jian-fu A session based deeplearning recommendation system considering userpurpose and time interval[J]. Computer Applications and Software, 2021, 38 (3): 190- 195
[14]   CAO L, PHILIP S Y Behavior informatics: an informatics perspective for behavior studies[J]. IEEE Intelligent Informatics Bulletin, 2009, 10 (1): 6- 11
[15]   WANG J, HUANG P, ZHAO H, et al. Billion-scale commodity embedding for e-commerce recommendation in Alibaba [C]// Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. London: ACM, 2018: 839-848.
[16]   PEROZZI B, AL-RFOU R, SKIENA S. Deepwalk: online learning of social representations [C]// Proceedings of the 20th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York: ACM, 2014: 701-710.
[17]   MIKOLOV T, CHEN K, CORRADO G, et al. Efficient estimation of word representations in vector space [EB/OL]. [2021-10-10]. https://arxiv.org/abs/1301.3781.
[18]   王永贵, 刘凯奇 一种优化聚类的协同过滤推荐算法[J]. 计算机工程与应用, 2020, (15): 66- 73
WANG Yong-gui, LIU Kai-qi Collaborative filtering recommendation algorithm for clustering optimization[J]. Computer Engineering and Applications, 2020, (15): 66- 73
doi: 10.3778/j.issn.1002-8331.1910-0095
[19]   岳希, 唐聃, 舒红平, 等 基于数据稀疏性的协同过滤推荐算法改进研究[J]. 工程科学与技术, 2020, 52 (1): 198- 202
YUE Xi, TANG Dan, SHU Hong-ping, et al Research on improvement of collaborative filtering recommendation algorithm based on data sparseness[J]. Advanced Engineering Sciences, 2020, 52 (1): 198- 202
[20]   WARSHAW P R, DAVIS F D Disentangling behavioral intention and behavioral expectation[J]. Journal of Experimental Social Psychology, 1985, 21 (3): 213- 228
doi: 10.1016/0022-1031(85)90017-4
[21]   AJZEN I The theory of planned behavior[J]. Organizational Behavior and Human Decision Processes, 1991, 50 (2): 179- 211
doi: 10.1016/0749-5978(91)90020-T
[22]   HAN J, PEI J, MORTAZAVI-ASL B, et al. Prefixspan: mining sequential patterns efficiently by prefix-projected pattern growth [C]// Proceedings of the 17th International Conference on Data Engineering. Heidelberg: IEEE, 2001: 215-224.
[23]   MovieLens 1M dataset [DB/OL]. [2021-10-10]. https://grouplens.org/datasets/movielens/1m/.
[24]   HARPER F M, KONSTAN J A The movielens datasets: history and context[J]. ACM Transactions on Interactive Intelligent Systems, 2015, 5 (4): 1- 19
[25]   RecSys2015 [DB/OL]. [2021-10-10]. https://recsys.acm.org/recs ys15/.
[26]   MCAULEY J. Amazon product data [DB/OL]. [2021-10-10]. http://jmcauley.ucsd.edu/data/amazon/.
[27]   MCAULEY J, TARGETT C, SHI Q, et al. Image-based recommendations on styles and substitutes [C]// Proceedings of the 38th International ACM SIGIR Conference on Research and Development in Information Retrieval. Santiago: ACM, 2015: 43-52.
[28]   YUAN Q, CONG G, SUN A. Graph-based point-of-interest recommendation with geographical and temporal influences [C]// Proceedings of the 23rd ACM International Conference on Conference on Information and Knowledge Management. Shanghai: ACM, 2014: 659-668.
[29]   ZHAO S L, ZHAO T, YANG H Q, et al. Stellar: spatial-temporal latent ranking for successive point-of-interest recommendation [C]// Proceedings of the 30th AAAI Conference on Artificial Intelligence. Phoenix: AAAI Press, 2016: 315–321.
[30]   LINDEN G, SMITH B, YORK J Amazon. com recommendations: item-to-item collaborative filtering[J]. IEEE Internet Computing, 2003, 7 (1): 76- 80
doi: 10.1109/MIC.2003.1167344
[31]   RENDLE S, FREUDENTHALER C, GANTNER Z, et al. BPR: Bayesian personalized ranking from implicit feedback [EB/OL]. [2021-10-10]. https://arxiv.org/abs/1205.2618.
[32]   LI J, REN P, CHEN Z, et al. Neural attentive session-based recommendation [C]// Proceedings of the 2017 ACM on Conference on Information and Knowledge Management. Singapore: ACM, 2017: 1419–1428.
[1] Si CHEN,Xiao-dong CAI,Zhen-zhen HOU,Bo LI. Aggregate graph embedding method based on non-uniform neighbor nodes sampling[J]. Journal of ZheJiang University (Engineering Science), 2019, 53(11): 2163-2167.