《计算机应用》唯一官方网站 ›› 2021, Vol. 41 ›› Issue (11): 3171-3177.DOI: 10.11772/j.issn.1001-9081.2021010047
所属专题: 人工智能
收稿日期:2021-01-12
									
				
											修回日期:2021-04-22
									
				
											接受日期:2021-04-29
									
				
											发布日期:2021-05-07
									
				
											出版日期:2021-11-10
									
				
			通讯作者:
					杨茂林
							作者简介:高铭蔚(1996—),男,河南南阳人,硕士研究生,主要研究方向:推荐系统、深度学习
        
                                                                                                            Mingwei GAO, Nan SANG, Maolin YANG( )
)
			  
			
			
			
                
        
    
Received:2021-01-12
									
				
											Revised:2021-04-22
									
				
											Accepted:2021-04-29
									
				
											Online:2021-05-07
									
				
											Published:2021-11-10
									
			Contact:
					Maolin YANG   
							About author:GAO Mingwei,born in 1996,M. S. candidate. His research
interests include recommendation system,deep learning摘要:
在交互式网络电视(IPTV)应用中,家庭电视终端往往由多名家庭成员共用,现有推荐算法难以从终端历史数据中分析出家庭成员的不同兴趣偏好。为了满足同一终端下不同成员的视频点播需求,提出了一种基于胶囊网络的IPTV视频点播推荐模型CapIPTV。首先,设计了一种基于胶囊网络路由机制的用户兴趣生成层,将终端历史行为数据作为输入,并通过胶囊网络的聚类特性得到不同家庭成员的兴趣表达;其次,利用注意力机制给不同的兴趣表达动态分配注意力权重;最后,提取出不同家庭成员的兴趣向量和点播视频的表示向量,计算两者内积后得出Top-N偏好推荐。在公开数据集MovieLens和真实广电数据集IPTV上的实验结果表明,CapIPTV的命中率(HR)、召回率(Recall)和归一化折损累计增益(DNCG)优于其他五种同类推荐模型。
中图分类号:
高铭蔚, 桑楠, 杨茂林. 基于胶囊网络的交互式网络电视视频点播推荐模型[J]. 计算机应用, 2021, 41(11): 3171-3177.
Mingwei GAO, Nan SANG, Maolin YANG. IPTV video-on-demand recommendation model based on capsule network[J]. Journal of Computer Applications, 2021, 41(11): 3171-3177.
| 符号 | 含义 | 
|---|---|
| IPTV终端 | |
| T | IPTV终端集 | 
| 点播视频:电影、电视剧等 | |
| 点播视频资源池 | |
| 终端 | |
| 终端 | |
| 点播视频 | |
| 终端兴趣矩阵所包含兴趣向量的数量 | 
表1 IPTV推荐模型符号定义
Tab. 1 Symbol definition for IPTV recommendation model
| 符号 | 含义 | 
|---|---|
| IPTV终端 | |
| T | IPTV终端集 | 
| 点播视频:电影、电视剧等 | |
| 点播视频资源池 | |
| 终端 | |
| 终端 | |
| 点播视频 | |
| 终端兴趣矩阵所包含兴趣向量的数量 | 
| 数据集 | 用户数 | 视频数 | 交互数 | 稠密度 | 
|---|---|---|---|---|
| MovieLens | 138 493 | 18 345 | 19 984 024 | 0.007 86 | 
| IPTV | 642 809 | 20 148 | 22 977 923 | 0.001 77 | 
表2 实验数据集统计信息
Tab. 2 Statistics of experimental datasets
| 数据集 | 用户数 | 视频数 | 交互数 | 稠密度 | 
|---|---|---|---|---|
| MovieLens | 138 493 | 18 345 | 19 984 024 | 0.007 86 | 
| IPTV | 642 809 | 20 148 | 22 977 923 | 0.001 77 | 
| 超参数 | 值 | 
|---|---|
| 表示向量维度 | 64 | 
| 批处理大小 | 128 | 
| 学习率 | 0.001 | 
| 行为序列最大长度 | 20 | 
| 优化器 | Adam | 
表3 模型超参数设置
Tab. 3 Hyperparameter setting of model
| 超参数 | 值 | 
|---|---|
| 表示向量维度 | 64 | 
| 批处理大小 | 128 | 
| 学习率 | 0.001 | 
| 行为序列最大长度 | 20 | 
| 优化器 | Adam | 
| 模型 | 单次迭代训练 所消耗的时间 | 模型 | 单次迭代训练 所消耗的时间 | 
|---|---|---|---|
| YouTube DNN | 4 | MIND | 13 | 
| DIN | 5 | CapIPTV | 9 | 
| DMIN | 10 | 
表4 不同模型单次迭代训练时间对比 (s)
Tab. 4 Training time comparison of different models in single iteration
| 模型 | 单次迭代训练 所消耗的时间 | 模型 | 单次迭代训练 所消耗的时间 | 
|---|---|---|---|
| YouTube DNN | 4 | MIND | 13 | 
| DIN | 5 | CapIPTV | 9 | 
| DMIN | 10 | 
| K | MovieLens | IPTV | ||||
|---|---|---|---|---|---|---|
| Recall@50 | HR@50 | NDCG@50 | Recall@50 | HR@50 | NDCG@50 | |
| 1 | 28.85 | 88.92 | 60.79 | 76.34 | 92.43 | 87.54 | 
| 2 | 26.51 | 87.40 | 58.48 | 76.75 | 92.92 | 88.25 | 
| 3 | 25.54 | 87.19 | 57.78 | 77.01 | 93.24 | 88.71 | 
| 4 | 24.81 | 86.52 | 56.84 | 77.36 | 93.98 | 89.53 | 
| 5 | 24.35 | 86.57 | 56.43 | 77.01 | 93.46 | 89.00 | 
表5 不同参数K下的CapIPTV模型在MovieLens和IPTV数据集上的性能比较 (%)
Tab. 5 Performance comparison of CapIPTV model with different parameter K on MovieLens and IPTV datasets
| K | MovieLens | IPTV | ||||
|---|---|---|---|---|---|---|
| Recall@50 | HR@50 | NDCG@50 | Recall@50 | HR@50 | NDCG@50 | |
| 1 | 28.85 | 88.92 | 60.79 | 76.34 | 92.43 | 87.54 | 
| 2 | 26.51 | 87.40 | 58.48 | 76.75 | 92.92 | 88.25 | 
| 3 | 25.54 | 87.19 | 57.78 | 77.01 | 93.24 | 88.71 | 
| 4 | 24.81 | 86.52 | 56.84 | 77.36 | 93.98 | 89.53 | 
| 5 | 24.35 | 86.57 | 56.43 | 77.01 | 93.46 | 89.00 | 
| 模型 | Recall@50 | HR@50 | NDCG@50 | 
|---|---|---|---|
| ItemPop | 12.06 | 64.68 | 33.39 | 
| YouTube DNN | 27.13 | 87.83 | 59.28 | 
| DIN | 26.87 | 87.53 | 59.44 | 
| DMIN | 27.52 | 87.76 | 59.65 | 
| MIND | 27.89 | 87.97 | 59.73 | 
| CapIPTV | 28.85 | 88.92 | 60.79 | 
表6 六种推荐模型在MovieLens数据集上的性能比较 (%)
Tab. 6 Performance comparison of 6 recommendation models on MovieLens dataset
| 模型 | Recall@50 | HR@50 | NDCG@50 | 
|---|---|---|---|
| ItemPop | 12.06 | 64.68 | 33.39 | 
| YouTube DNN | 27.13 | 87.83 | 59.28 | 
| DIN | 26.87 | 87.53 | 59.44 | 
| DMIN | 27.52 | 87.76 | 59.65 | 
| MIND | 27.89 | 87.97 | 59.73 | 
| CapIPTV | 28.85 | 88.92 | 60.79 | 
| 模型 | Recall@50 | HR@50 | NDCG@50 | 
|---|---|---|---|
| ItemPop | 49.55 | 72.11 | 61.67 | 
| YouTube DNN | 75.24 | 91.44 | 87.36 | 
| DIN | 75.43 | 90.97 | 87.71 | 
| DMIN | 75.87 | 91.89 | 88.59 | 
| MIND | 76.29 | 92.67 | 88.73 | 
| CapIPTV | 77.36 | 93.98 | 89.53 | 
表7 六种推荐模型在IPTV数据集上的性能比较 (%)
Tab. 7 Performance comparison of 6 recommendation models on IPTV dataset
| 模型 | Recall@50 | HR@50 | NDCG@50 | 
|---|---|---|---|
| ItemPop | 49.55 | 72.11 | 61.67 | 
| YouTube DNN | 75.24 | 91.44 | 87.36 | 
| DIN | 75.43 | 90.97 | 87.71 | 
| DMIN | 75.87 | 91.89 | 88.59 | 
| MIND | 76.29 | 92.67 | 88.73 | 
| CapIPTV | 77.36 | 93.98 | 89.53 | 
| 模型 | Recall@50 | HR@50 | NDCG@50 | 
|---|---|---|---|
| CapIPTV_1 | 75.79 | 91.68 | 88.43 | 
| CapIPTV_2 | 76.43 | 92.99 | 88.52 | 
| CapIPTV_3 | 76.72 | 92.79 | 88.75 | 
| CapIPTV | 77.36 | 93.98 | 89.53 | 
表8 CapIPTV模型的消融实验结果 (%)
Tab. 8 Ablation experimental results of CapIPTV model
| 模型 | Recall@50 | HR@50 | NDCG@50 | 
|---|---|---|---|
| CapIPTV_1 | 75.79 | 91.68 | 88.43 | 
| CapIPTV_2 | 76.43 | 92.99 | 88.52 | 
| CapIPTV_3 | 76.72 | 92.79 | 88.75 | 
| CapIPTV | 77.36 | 93.98 | 89.53 | 
| 1 | ZHANG S, YAO L N, SUN A X, et al. Deep learning based recommender system: a survey and new perspectives [J]. ACM Computing Surveys, 2019, 52(1): Article No.5. 10.1145/3285029 | 
| 2 | DAS D, HORST H T. Recommender systems for TV: AAAI Technical Report WS-98-08-008 [R]. Eindhoven: Philips Research Laboratories Eindhoven, 1998: 1-2. | 
| 3 | YU Z W, ZHOU X S, HAO Y B, et al. TV program recommendation for multiple viewers based on user profile merging [J]. User Modeling and User-Adapted Interaction, 2006, 16(1):63-82. 10.1007/s11257-006-9005-6 | 
| 4 | KIM E, PYO S, PARK E, et al. An automatic recommendation scheme of TV program contents for (IP)TV personalization [J]. IEEE Transactions on Broadcasting, 2011, 57(3): 674-684. 10.1109/tbc.2011.2161409 | 
| 5 | SHIN H, LEE M, KIM E Y. Personalized digital TV content recommendation with integration of user behavior profiling and multimodal content rating [J]. IEEE Transactions on Consumer Electronics, 2009, 55(3): 1417-1423. 10.1109/tce.2009.5278008 | 
| 6 | TENG Y, HE L. Rating prediction algorithm and recommendation based on user behavior in IPTV [C]// Proceedings of the 2012 2nd International Conference on Consumer Electronics, Communications and Networks. Piscataway: IEEE, 2012: 3373-3378. 10.1109/cecnet.2012.6201480 | 
| 7 | LECUN Y, BENGIO Y, HINTON G. Deep learning [J]. Nature, 2015, 521(7553): 436-444. 10.1038/nature14539 | 
| 8 | BATMAZ Z, YUREKLI A, BILGE A, et al. A review on deep learning for recommender systems: challenges and remedies [J]. Artificial Intelligence Review, 2019, 52(1): 1-37. 10.1007/s10462-018-9654-y | 
| 9 | COVINGTON P, ADAMS J, SARGIN E. Deep neural networks for YouTube recommendations [C]// Proceedings of the 2016 10th ACM Conference on Recommender Systems. New York: ACM, 2016: 191-198. 10.1145/2959100.2959190 | 
| 10 | ZHOU G R, ZHU X Q, SONG C R, et al. Deep interest network for click-through rate prediction [C]// Proceedings of the 2018 24th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York: ACM, 2018: 1059-1068. 10.1145/3219819.3219823 | 
| 11 | XIAO Z B, YANG L W, JIANG W, et al. Deep multi-interest network for click-through rate prediction [C]// Proceedings of the 2020 29th ACM International Conference on Information and Knowledge Management. New York: ACM, 2020: 2265-2268. 10.1145/3340531.3412092 | 
| 12 | TANG J X, WANG K. Personalized Top-N sequential recommendation via convolutional sequence embedding [C]// Proceedings of the 2018 11th ACM International Conference on Web Search and Data Mining. New York: ACM, 2018: 565-573. 10.1145/3159652.3159656 | 
| 13 | HE X N, LIAO L Z, ZHANG H W, et al. Neural collaborative filtering [C]// Proceedings of the 2017 26th International Conference on World Wide Web. New York: ACM, 2017: 173-182. 10.1145/3038912.3052569 | 
| 14 | 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 2017 26th International Joint Conference on Artificial Intelligence. New York: ACM, 2017: 1725-1731. 10.24963/ijcai.2017/239 | 
| 15 | XUE H J, DAI X Y, ZHANG J B, et al. Deep matrix factorization models for recommender systems [C]// Proceedings of the 2017 26th International Joint Conference on Artificial Intelligence. New York: ACM, 2017: 3203-3209. 10.24963/ijcai.2017/447 | 
| 16 | HINTON G E, KRIZHEVSKY A, WANG S D. Transforming auto-encoders [C]// Proceedings of the 2011 International Conference on Artificial Neural Networks, LNCS6791. Berlin: Springer, 2011: 44-51. | 
| 17 | SABOUR S, FROSST N, HINTON G E. Dynamic routing between capsules [C]// Proceedings of the 2017 31st Conference on Neural Information Processing Systems. Red Hook: Curran Associates Inc., 2017: 3859-3869. | 
| 18 | LI C, LIU Z Y, WU M M, et al. Multi-interest network with dynamic routing for recommendation at Tmall [C]// Proceedings of the 2019 28th ACM International Conference on Information and Knowledge Management. New York: ACM, 2019: 2615-2623. 10.1145/3357384.3357814 | 
| 19 | LI C L, QUAN C, PENG L, et al. A capsule network for recommendation and explaining what you like and dislike [C]// Proceedings of the 2019 42nd International ACM SIGIR Conference on Research and Development in Information Retrieval. New York: ACM, 2019: 275-284. 10.1145/3331184.3331216 | 
| 20 | JEAN S, CHO K, MEMISEVIC R, et al. On using very large target vocabulary for neural machine translation [C]// Proceedings of the 2015 53rd Annual Meeting of the Association for Computational Linguistics/ the 7th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). Stroudsburg: ACL, 2015: 1-10. 10.3115/v1/p15-1001 | 
| 21 | JOHNSON J, DOUZE M, JÉGOU H. Billion-scale similarity search with GPUs [J]. IEEE Transactions on Big Data, 2021, 7(3): 535-547. 10.1109/tbdata.2019.2921572 | 
| 22 | LIANG D W, KRISHNAN R G, HOFFMAN M D, et al. Variational autoencoders for collaborative filtering [C]// Proceedings of the 2018 World Wide Web Conference. New York: ACM, 2018: 689-698. 10.1145/3178876.3186150 | 
| 23 | MA J X, ZHOU C, CUI P, et al. Learning disentangled representations for recommendation [C]// Proceedings of the 2019 33rd Conference on Neural Information Processing Systems. New York: ACM, 2019: 5712-5723. 10.1145/3394486.3403091 | 
| 24 | CREMONESI P, KOREN Y, TURRIN R. Performance of recommender algorithms on top-N recommendation tasks [C]// Proceedings of the 2010 4th ACM Conference on Recommender Systems. New York: ACM, 2010: 39-46. 10.1145/1864708.1864721 | 
| 25 | 南宁,杨程屹,武志昊.基于多图神经网络的会话感知推荐模型[J].计算机应用,2021,41(2):330-336. 10.1016/j.future.2021.06.007 | 
| NAN N, YANG C Y, WU Z H. Multi-graph neural network-based session perception recommendation model [J] Journal of Computer Application, 2021, 41(2): 330-336. 10.1016/j.future.2021.06.007 | 
| [1] | 唐廷杰, 黄佳进, 秦进. 基于图辅助学习的会话推荐[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2711-2718. | 
| [2] | 赵志强, 马培红, 黑新宏. 基于双重注意力机制的人群计数方法[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2886-2892. | 
| [3] | 秦璟, 秦志光, 李发礼, 彭悦恒. 基于概率稀疏自注意力神经网络的重性抑郁疾患诊断[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2970-2974. | 
| [4] | 李力铤, 华蓓, 贺若舟, 徐况. 基于解耦注意力机制的多变量时序预测模型[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2732-2738. | 
| [5] | 薛凯鹏, 徐涛, 廖春节. 融合自监督和多层交叉注意力的多模态情感分析网络[J]. 《计算机应用》唯一官方网站, 2024, 44(8): 2387-2392. | 
| [6] | 汪雨晴, 朱广丽, 段文杰, 李书羽, 周若彤. 基于交互注意力机制的心理咨询文本情感分类模型[J]. 《计算机应用》唯一官方网站, 2024, 44(8): 2393-2399. | 
| [7] | 高鹏淇, 黄鹤鸣, 樊永红. 融合坐标与多头注意力机制的交互语音情感识别[J]. 《计算机应用》唯一官方网站, 2024, 44(8): 2400-2406. | 
| [8] | 李钟华, 白云起, 王雪津, 黄雷雷, 林初俊, 廖诗宇. 基于图像增强的低照度人脸检测[J]. 《计算机应用》唯一官方网站, 2024, 44(8): 2588-2594. | 
| [9] | 莫尚斌, 王文君, 董凌, 高盛祥, 余正涛. 基于多路信息聚合协同解码的单通道语音增强[J]. 《计算机应用》唯一官方网站, 2024, 44(8): 2611-2617. | 
| [10] | 唐廷杰, 黄佳进, 秦进, 陆辉. 基于图共现增强多层感知机的会话推荐[J]. 《计算机应用》唯一官方网站, 2024, 44(8): 2357-2364. | 
| [11] | 熊武, 曹从军, 宋雪芳, 邵云龙, 王旭升. 基于多尺度混合域注意力机制的笔迹鉴别方法[J]. 《计算机应用》唯一官方网站, 2024, 44(7): 2225-2232. | 
| [12] | 李欢欢, 黄添强, 丁雪梅, 罗海峰, 黄丽清. 基于多尺度时空图卷积网络的交通出行需求预测[J]. 《计算机应用》唯一官方网站, 2024, 44(7): 2065-2072. | 
| [13] | 毛典辉, 李学博, 刘峻岭, 张登辉, 颜文婧. 基于并行异构图和序列注意力机制的中文实体关系抽取模型[J]. 《计算机应用》唯一官方网站, 2024, 44(7): 2018-2025. | 
| [14] | 刘丽, 侯海金, 王安红, 张涛. 基于多尺度注意力的生成式信息隐藏算法[J]. 《计算机应用》唯一官方网站, 2024, 44(7): 2102-2109. | 
| [15] | 徐松, 张文博, 王一帆. 基于时空信息的轻量视频显著性目标检测网络[J]. 《计算机应用》唯一官方网站, 2024, 44(7): 2192-2199. | 
| 阅读次数 | ||||||
| 全文 |  | |||||
| 摘要 |  | |||||