基于自适应群组重排的长尾推荐模型

doi:10.11772/j.issn.1001-9081.2022030455

《计算机应用》唯一官方网站 ›› 2023, Vol. 43 ›› Issue (4): 1122-1128.DOI: 10.11772/j.issn.1001-9081.2022030455

• 数据科学与技术 • 上一篇

基于自适应群组重排的长尾推荐模型

金苍宏¹^,²(), 邵育华², 何琴芳²

^1.浙大城市学院城市大脑研究院，杭州310015
^2.浙大城市学院计算机与计算科学学院，杭州310015

收稿日期:2022-04-11 修回日期:2022-08-08 接受日期:2022-08-15 发布日期:2023-01-11 出版日期:2023-04-10
通讯作者: 金苍宏
作者简介:邵育华（2001—），男，浙江杭州人，主要研究方向：推荐系统；
何琴芳（1999—），女，浙江嘉兴人，主要研究方向：大数据、人工智能。
基金资助:
浙江省重点研发计划项目(2021C02060);浙江省自然科学基金资助项目(LY21F020003);浙大城市学院科研培育基金资助课题(X?202206)

Long-tail recommendation model based on adaptive group reranking

Canghong JIN¹^,²(), Yuhua SHAO², Qinfang HE²

^1.City Brain Institute，Zhejiang University City College，Hangzhou Zhejiang 310015，China
^2.School of Computer and Computing Science，Zhejiang University City College，Hangzhou Zhejiang 310015，China

Received:2022-04-11 Revised:2022-08-08 Accepted:2022-08-15 Online:2023-01-11 Published:2023-04-10
Contact: Canghong JIN
About author:SHAO Yuhua， born in 2001. His research interests include recommender system.
HE Qinfang， born in 1999. Her research interests include big data， artificial intelligence.
Supported by:
Zhejiang Key Science and Technology Program(2021C02060);Natural Science Foundation of Zhejiang Province(LY21F020003);Scientific Development Foundation of Zhejiang University City College(X-202206)

摘要/Abstract

摘要：

针对传统推荐算法过度关注推荐的精度而导致的长尾问题，即热门项目拥有过高的推荐量的同时非热门项目长时间不被关注，提出一种基于欧氏距离构建二维加权相似度并融入自适应群组重排的多目标优化推荐模型（MDOM）——自适应群组重排的推荐模型（AGRM）。首先，利用欧氏距离构建二维加权相似度度量，根据个体历史行为记录动态设定替换比例，并利用融入群组的多目标优化算法解决长尾推荐问题；其次，设计两个简明的目标函数，并同时考虑流行度和长尾关注度，以降低目标函数的复杂性；然后，基于二维加权相似度度量，选择用户子集作为“最佳推荐用户组”，并计算帕累托最优解。在MovieLens 1M和Yahoo数据集上的实验结果表明，AGRM的覆盖率表现最优，与基于物品相似的协同过滤（ItemCF）算法相比，分别平均提升了4.11、25.38个百分点；与用于Top-N推荐的具有浅并行路径的深度变分自动编码器（VASP）模型相比，分别平均提升了8.38、33.19个百分点。在Yahoo数据集上，AGRM的推荐的平均流行度最低，表明AGRM能够推荐更多长尾项目。

Abstract:

The traditional recommendation algorithms pay too much attention to the precision of recommendation， which leads to the high recommendation rate of popular items. At the same time， the unpopular items are not paid attention to for a long time. This is a classic long-tail problem. In response to this problem， a Multi-objective Dimension Optimization recommendation Model （MDOM）， named Adaptive Group Reranking recommendation Model （AGRM） was proposed， with the construction of two-dimensional weighted similarity based on Euclidean distance and the incorporation of adaptive group reranking. Firstly， a two-dimensional weighted similarity measure was constructed using Euclidean distance， the replacement ratio was set dynamically according to the individual’s historical behavior records， and the long-tail recommendation problem was solved by using the multi-objective optimization algorithm integrated with group. Secondly， two concise objective functions were designed， and the complexity of the objective functions was reduced by taking popularity and long-tail attention into account. Thirdly， based on the two-dimensional weighted similarity measure， a user subset was selected as the "best recommended user group"， and the Pareto optimal solution was calculated. Experimental results on MovieLens 1M and Yahoo datasets show that the coverage of AGRM is the best， with an average increase of 4.11 percentage points and 25.38 percentage points respectively compared to that of Item-based Collaborative Filtering （ItemCF） algorithm， and an average increase of 8.38 percentage points and 33.19 percentage points respectively compared to that of Deep Variational Autoencoder with Shallow Parallel Path for Top-N Recommendation （VASP） model. On Yahoo dataset， the average popularity of AGRM recommendation is the lowest， indicating that AGRM can recommend more long-tail items.

Key words: long-tail recommendation, adaptive reranking, multi-objective optimization, effective user group, behavior decision model

中图分类号:

TP391.4

金苍宏, 邵育华, 何琴芳. 基于自适应群组重排的长尾推荐模型[J]. 计算机应用, 2023, 43(4): 1122-1128.

Canghong JIN, Yuhua SHAO, Qinfang HE. Long-tail recommendation model based on adaptive group reranking[J]. Journal of Computer Applications, 2023, 43(4): 1122-1128.

图/表 4

参考文献 35

1	YANG X W， GUO Y， LIU Y， et al. A survey of collaborative filtering based social recommender systems［J］. Computer Communications， 2014， 41： 1-10. 10.1016/j.comcom.2013.06.009
2	ADOMAVICIUS G， TUZHILIN A. Toward the next generation of recommender systems： a survey of the state-of-the-art and possible extensions［J］. IEEE Transactions on Knowledge and Data Engineering， 2005， 17（6）： 734-749. 10.1109/tkde.2005.99
3	RANJBAR KERMANY N， ALIZADEH S H. A hybrid multi-criteria recommender system using ontology and neuro-fuzzy techniques［J］. Electronic Commerce Research and Applications， 2017， 21： 50-64. 10.1016/j.elerap.2016.12.005
4	RANJBAR KERMANY N， ZHAO W L， YANG J， et al. ReInCre： enhancing collaborative filtering recommendations by incorporating user rating credibility［C］// Proceedings of the 2020 International Conference on Web Information Systems Engineering， CCIS 1155. Singapore： Springer， 2020： 64-72.
5	骆锦潍，刘杜钢，潘微科，等. 基于改进的倾向得分估计的无偏推荐模型［J］. 计算机应用， 2021， 41（12）： 3508-3514. 10.11772/j.issn.1001-9081.2021060910
	LUO J W， LIU D G， PAN W K， et al. Unbiased recommendation model based on improved propensity score estimation［J］. Journal of Computer Applications， 2021， 41（12）： 3508-3514. 10.11772/j.issn.1001-9081.2021060910
6	PARK Y J， TUZHILIN A. The long tail of recommender systems and how to leverage it［C］// Proceedings of the 2008 ACM Conference on Recommender Systems. New York： ACM， 2008： 11-18. 10.1145/1454008.1454012
7	KIM Y， KIM K， PARK C， et al. Sequential and diverse recommendation with long tail［C］// Proceedings of the 28th International Joint Conference on Artificial Intelligence. California： ijcai.org， 2019： 2740-2746. 10.24963/ijcai.2019/380
8	PANG J N， GUO J， ZHANG W. Using multi-objective optimization to solve the long tail problem in recommender system［C］// Proceedings of the 2019 Pacific-Asia Conference on Knowledge Discovery and Data Mining， LNCS 11441. Cham： Springer， 2019： 302-313.
9	WANG S F， GONG M G， LI H L， et al. Multi-objective optimization for long tail recommendation［J］. Knowledge-Based Systems， 2016， 104： 145-155. 10.1016/j.knosys.2016.04.018
10	RIBEIRO M T， LACERDA A， VELOSO A， et al. Pareto-efficient hybridization for multi-objective recommender systems［C］// Proceedings of the 6th ACM Conference on Recommender Systems. New York： ACM， 2012： 19-26. 10.1145/2365952.2365962
11	YIN H Z， CUI B， LI J.， et al. Challenging the long tail recommendation［J］. Proceedings of the VLDB Endowment， 2012， 5（9）： 896-907. 10.14778/2311906.2311916
12	JOHNSON J， NG Y K. Enhancing long tail item recommendations using tripartite graphs and Markov process［C］// Proceedings of the 2017 IEEE/WIC/ACM International Conference on Web Intelligence. New York： ACM， 2017： 761-768. 10.1145/3106426.3106439
13	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. 10.11897/SP.J.1016.2018.01619
	HUANG L W， JIANG B T， LYU S Y， et al. Survey on deep learning based recommender systems［J］. Chinese Journal of Computers， 2018， 41（7）： 1619-1647. 10.11897/SP.J.1016.2018.01619
14	ADOMAVICIUS G， KWON Y. Improving aggregate recommendation diversity using ranking-based techniques［J］. IEEE Transactions on Knowledge and Data Engineering， 2012， 24（5）： 896-911. 10.1109/tkde.2011.15
15	MIYAMOTO S， ZAMAMI T， YAMANA H. Improving recommendation diversity across users by reducing frequently recommended items［C］// Proceedings of the 2018 IEEE International Conference on Big Data. Piscataway： IEEE， 2018： 5392-5394. 10.1109/bigdata.2018.8622314
16	ABDOLLAHPOURI H， BURKE R， MOBASHER B. Managing popularity bias in recommender systems with personalized re-ranking［EB/OL］. （2019-08-12）［2022-02-23］.. 10.1145/3383313.3418487
17	KRISHNAN A， SHARMA A， SANKAR A， et al. An adversarial approach to improve long-tail performance in neural collaborative filtering［C］// Proceedings of the 27th ACM International Conference on Information and Knowledge Management. New York： ACM， 2018： 1491-1494. 10.1145/3269206.3269264
18	HE X N， LIAO L Z， ZHANG H W， et al. Neural collaborative filtering［C］// Proceedings of the 26th International Conference on World Wide Web. Republic and Canton of Geneva： International World Wide Web Conferences Steering Committee， 2017： 173-182. 10.1145/3038912.3052569
19	OUYANG W T， ZHANG X W， REN S K， et al. Learning graph meta embeddings for cold-start ads in click-through rate prediction［C］// Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval. New York： ACM， 2021： 1157-1166. 10.1145/3404835.3462879
20	CHEN Y， HOU Y Q， LENG S， et al. Long-tail hashing［C］// Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval. New York： ACM， 2021： 1328-1338. 10.1145/3404835.3462888
21	MA J Q， ZHAO Z， YI X Y， et al. Modeling task relationships in multi-task learning with multi-gate mixture-of-experts［C］// Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York： ACM， 2018： 1930-1939. 10.1145/3219819.3220007
22	YI X Y， YANG J， HONG L C， et al. Sampling-bias-corrected neural modeling for large corpus item recommendations［C］// Proceedings of the 13th ACM Conference on Recommender Systems. New York： ACM， 2019： 269-277. 10.1145/3298689.3346996
23	VANČURA V， KORDÍK P. Deep Variational Autoencoder with Shallow Parallel path for Top-N recommendation （VASP）［C］// Proceedings of the 2021 International Conference on Artificial Neural Networks， LNTCS 12895. Cham： Springer， 2021： 138-149.
24	ZHENG Y， WANG D X. A survey of recommender systems with multi-objective optimization［J］. Neurocomputing， 2022， 474： 141-153. 10.1016/j.neucom.2021.11.041
25	CACHEDA F， CARNEIRO V， FERNÁNDEZ D， et al. Comparison of collaborative filtering algorithms： limitations of current techniques and proposals for scalable， high-performance recommender systems［J］. ACM Transactions on the Web， 2011， 5（1）： No.2. 10.1145/1921591.1921593
26	HAN J Y， ZHENG L， HUANG H， et al. Deep latent factor model with hierarchical similarity measure for recommender systems［J］. Information Sciences， 2019， 503： 521-532. 10.1016/j.ins.2019.07.024
27	WU Y， DuBOIS C， ZHENG A X， et al. Collaborative denoising auto-encoders for top-N recommender systems［C］// Proceedings of the 9th ACM International Conference on Web Search and Data Mining. New York： ACM， 2016： 153-162. 10.1145/2835776.2835837
28	浙大城市学院. 一种基于提取有效多目标群组来解决长尾推荐问题的方法： 202210363831.X［P］. 2022-05-17.
	Zhejiang University City College. A method to solve the long tail recommendation problem based on extracting effective multi-objective groups： 202210363831.X［P］. 2022-05-17.
29	KRAMER O. Genetic algorithms［M］// Genetic Algorithm Essentials， SCI 679. Cham： Springer， 2017： 11-19. 10.1007/978-3-319-52156-5_2
30	SRINIVAS N， DEB K. Muiltiobjective optimization using nondominated sorting in genetic algorithms［J］. Evolutionary Computation， 1994， 2（3）： 221-248. 10.1162/evco.1994.2.3.221
31	ZHANG Q F， LI H. MOEA/D： a multiobjective evolutionary algorithm based on decomposition［J］. IEEE Transactions on Evolutionary Computation， 2007， 11（6）： 712-731. 10.1109/tevc.2007.892759
32	DEB K， PRATAP A， AGARWAL S， et al. A fast and elitist multiobjective genetic algorithm： NSGA-Ⅱ［J］. IEEE Transactions on Evolutionary Computation， 2002， 6（2）： 182-197. 10.1109/4235.996017
33	秦婧，张青博，王斌. 关注长尾物品的推荐方法［J］. 计算机应用， 2020， 40（2）： 454-458. 10.11772/j.issn.1001-9081.2019091665
	QIN Q， ZHANG Q B， WANG B. Recommendation method with focus on long tail items［J］. Journal of Computer Applications， 2020， 40（2）： 454-458. 10.11772/j.issn.1001-9081.2019091665
34	MovieLens 1 M dataset［DS/OL］. ［2022-02-23］..
35	Yahoo Webscope datasest［DS/OL］. ［2022-02-23］.. 10.1145/3416921.3416934

模型	N	MovieLens 1M		Yahoo		模型	N	MovieLens 1M		Yahoo
模型	N	精度/%	覆盖率/%	精度/%	覆盖率/%	模型	N	精度/%	覆盖率/%	精度/%	覆盖率/%
ItemCF^［25］	5	43.20	2.96	25.20	3.94	CDAE^［27］	5	39.07	14.60	40.01	4.41
	10	39.20	6.43	20.50	6.84		10	34.34	18.97	36.56	7.07
	20	31.50	12.45	16.35	17.63		20	29.96	23.73	31.78	11.47
	40	25.60	25.51	12.07	46.64		40	25.05	32.76	26.97	18.19
	80	19.14	49.90	7.85	82.02		80	19.58	43.06	21.18	29.01
LFM^［26］	5	22.00	0.66	22.08	1.51	VASP^［23］	5	33.23	6.23	30.13	5.10
	10	19.90	1.17	18.55	2.44		10	29.65	9.76	27.01	8.90
	20	17.05	1.94	13.75	3.94		20	23.57	13.63	22.67	13.12
	40	13.75	3.37	10.22	6.84		40	19.60	20.79	18.98	16.67
	80	10.25	5.87	6.93	12.88		80	15.66	25.49	14.80	22.28
UserCF^［25］	5	43.80	8.69	23.80	8.15	MDOM^［28］	5	23.29	4.86	15.60	15.66
	10	39.00	13.56	19.00	16.20		10	18.52	8.95	12.06	28.19
	20	33.60	22.57	15.25	26.17		20	15.34	16.51	10.08	50.35
	40	25.77	33.93	10.78	54.06		40	12.20	28.38	7.35	85.03
	80	19.29	52.09	7.10	87.57		80	8.71	57.99	5.45	―
NSGA-Ⅱ-CF-RS^［8］	5	17.21	4.13	10.84	15.18	AGRM	5	21.80	6.56	15.80	16.19
	10	15.25	6.67	7.14	25.83		10	18.50	9.36	13.40	27.84
	20	15.61	13.78	6.39	37.24		20	16.60	15.47	10.75	49.37
	40	12.17	24.09	5.36	66.82		40	12.17	28.17	7.95	83.16
	80	8.58	41.91	3.44	―		80	8.56	58.24	5.50	―

模型	N	MovieLens 1M		Yahoo		模型	N	MovieLens 1M		Yahoo
模型	N	精度/%	覆盖率/%	精度/%	覆盖率/%	模型	N	精度/%	覆盖率/%	精度/%	覆盖率/%
ItemCF^［25］	5	43.20	2.96	25.20	3.94	CDAE^［27］	5	39.07	14.60	40.01	4.41
	10	39.20	6.43	20.50	6.84		10	34.34	18.97	36.56	7.07
	20	31.50	12.45	16.35	17.63		20	29.96	23.73	31.78	11.47
	40	25.60	25.51	12.07	46.64		40	25.05	32.76	26.97	18.19
	80	19.14	49.90	7.85	82.02		80	19.58	43.06	21.18	29.01
LFM^［26］	5	22.00	0.66	22.08	1.51	VASP^［23］	5	33.23	6.23	30.13	5.10
	10	19.90	1.17	18.55	2.44		10	29.65	9.76	27.01	8.90
	20	17.05	1.94	13.75	3.94		20	23.57	13.63	22.67	13.12
	40	13.75	3.37	10.22	6.84		40	19.60	20.79	18.98	16.67
	80	10.25	5.87	6.93	12.88		80	15.66	25.49	14.80	22.28
UserCF^［25］	5	43.80	8.69	23.80	8.15	MDOM^［28］	5	23.29	4.86	15.60	15.66
	10	39.00	13.56	19.00	16.20		10	18.52	8.95	12.06	28.19
	20	33.60	22.57	15.25	26.17		20	15.34	16.51	10.08	50.35
	40	25.77	33.93	10.78	54.06		40	12.20	28.38	7.35	85.03
	80	19.29	52.09	7.10	87.57		80	8.71	57.99	5.45	―
NSGA-Ⅱ-CF-RS^［8］	5	17.21	4.13	10.84	15.18	AGRM	5	21.80	6.56	15.80	16.19
	10	15.25	6.67	7.14	25.83		10	18.50	9.36	13.40	27.84
	20	15.61	13.78	6.39	37.24		20	16.60	15.47	10.75	49.37
	40	12.17	24.09	5.36	66.82		40	12.17	28.17	7.95	83.16
	80	8.58	41.91	3.44	―		80	8.56	58.24	5.50	―

[1]	李兴佳, 杨秋辉, 洪玫, 潘春霞, 刘瑞航. 基于历史数据和多目标优化的测试用例排序方法[J]. 《计算机应用》唯一官方网站, 2023, 43(1): 221-226.
[2]	马艳芳, 张文, 李宗敏, 闫芳, 郭凌云. 考虑负效应的垃圾回收两级选址‒路径模型与算法[J]. 《计算机应用》唯一官方网站, 2023, 43(1): 289-298.
[3]	张翔宇, 杨阳, 冯国徽, 秦川. 基于多目标优化的加密图像可逆信息隐藏[J]. 《计算机应用》唯一官方网站, 2022, 42(6): 1716-1723.
[4]	张金泉, 徐寿伟, 李信诚, 王重洋, 徐景芝. 基于正交自适应鲸鱼优化的云计算任务调度[J]. 《计算机应用》唯一官方网站, 2022, 42(5): 1516-1523.
[5]	张京, 朱爱红. 基于遗传算法和粒子群优化的列车自动驾驶速度曲线优化方法[J]. 《计算机应用》唯一官方网站, 2022, 42(2): 599-605.
[6]	郭潇, 李春山, 张宇跃, 初佃辉. 基于自适应多目标强化学习的服务集成方法[J]. 《计算机应用》唯一官方网站, 2022, 42(11): 3500-3505.
[7]	罗明宇, 骆晓萌, 朱文敏, 张磊, 范秀敏. 船舶狭小空间虚拟人维修姿态建模技术[J]. 计算机应用, 2021, 41(8): 2466-2472.
[8]	张闻强, 邢征, 杨卫东. 基于多区域采样策略的混合粒子群优化求解多目标柔性作业车间调度问题[J]. 计算机应用, 2021, 41(8): 2249-2257.
[9]	张祥飞, 鲁宇明, 张平生. 基于协同进化的约束多目标优化算法[J]. 计算机应用, 2021, 41(7): 2012-2018.
[10]	马晓航, 廖灵霞, 李智, 秦斌, 赵涵捷. 基于动态混合超时的软件定义网络多目标优化[J]. 《计算机应用》唯一官方网站, 2021, 41(12): 3658-3665.
[11]	乔钢柱, 王瑞, 孙超利. 基于分解的高维多目标改进进化算法[J]. 《计算机应用》唯一官方网站, 2021, 41(11): 3097-3103.
[12]	李二超, 李康伟. 基于最小距离和聚合策略的分解多目标进化算法[J]. 计算机应用, 2021, 41(1): 22-28.
[13]	李二超, 杨蓉蓉. 基于自适应反向学习的多目标分布估计算法[J]. 计算机应用, 2021, 41(1): 15-21.
[14]	王博, 刘连生, 韩绍程, 祝世兴. 基于多策略融合的混合多目标蝗虫优化算法[J]. 计算机应用, 2020, 40(9): 2670-2676.
[15]	张凯, 周德云, 杨振, 潘潜. 多约束集群目标下武器-目标分配问题[J]. 计算机应用, 2020, 40(3): 902-911.

基于自适应群组重排的长尾推荐模型

Long-tail recommendation model based on adaptive group reranking

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 4

参考文献 35

相关文章 15

编辑推荐

Metrics