Multi-action Click Prediction Model for Short Video Users Based On User’s Behavior Sequence
-
摘要: 目前主流的点击预测模型采用线性模型和深度神经网络相结合的方法学习用户与物品之间特征交互，忽略了用户的历史行为本质上是一个动态序列的事实，从而导致无法有效捕获用户行为序列中蕴含的时间信息。为此，该文提出了基于用户行为序列的短视频用户多行为点击预测模型(USCP)。该模型将用户的历史行为按交互时间的顺序排序，生成用户历史行为序列。在DeepFM模型的基础上引入词嵌入模型Word2Vec，根据用户历史行为序列自适应学习到该用户的动态兴趣，有效捕获到用户兴趣的变化。在某短视频平台上公开的脱敏数据集上进行了对比实验，评价指标采用GAUC(Group AUC)，结果表明该模型性能优于其他几个模型。Abstract: At present, the mainstream click prediction model uses the combination of linear model and deep neural network to learn the characteristic interaction between users and items, ignoring the fact that the user’s historical behavior is essentially a dynamic sequence, resulting in the inability to capture effectively the time information contained in the user’s behavior sequence. Therefore, a short video USer multi behavior Click Prediction model (USCP) based on user behavior sequence is proposed in this paper. The model sorts the user’s historical behavior in the order of interaction time, and generates the user’s historical behavior sequence. Based on the DeepFM model, the word embedding model word2vec is introduced to learn adaptively the user’s dynamic interest according to the user’s historical behavior sequence and capture effectively the changes of user interest. A comparative experiment is carried out on the desensitization data set published on a short video platform. The evaluation index adopts GAUC (Group AUC). The results show that the performance of this model is better than other models.
-
Key words:
- Behavior sequence /
- Deep learning /
- DeepFM /
- Click prediction /
- Word2Vec
-
表 2 模型性能比较
模型 查看评论 点赞 点击头像 转发 GAUC Wide&Deep 0.61542 0.61575 0.69122 0.65389 0.63452 MMOE 0.63147 0.61776 0.71244 0.68328 0.64873 DeepFM 0.63176 0.61734 0.71939 0.70830 0.65261 USCP 0.63561 0.63183 0.72806 0.72445 0.66185 
下载: 导出CSV
-
[1] WANG Xinfei. A survey of online advertising click-through rate prediction models[C]. 2020 IEEE International Conference on Information Technology, Big Data and Artificial Intelligence (ICIBA), Chongqing, China, 2020: 516–521. [2] CHAPELLE O, MANAVOGLU E, and ROSALES R. Simple and scalable response prediction for display advertising[J]. ACM Transactions on Intelligent Systems and Technology, 2015, 5(4): 61. doi: 10.1145/2532128 [3] GAI Kun, ZHU Xiaoqiang, LI Han, et al. Learning piece-wise linear models from large scale data for Ad click prediction[J]. arXiv: 1704.05194, 2017. [4] RENDLE S. Factorization machines[C]. 2010 IEEE International Conference on Data Mining, Sydney, Australia, 2010: 995–1000. [5] CHENG H T, KOC L, HARMSEN J, et al. Wide & deep learning for recommender systems[C]. The 1st Workshop on Deep Learning for Recommender Systems, Boston, USA, 2016: 7–10. [6] GUO Huifeng, TANG Ruiming, YE Yunming, et al. DeepFM: A factorization-machine based neural network for CTR prediction[C]. The 26th International Joint Conference on Artificial Intelligence, Melbourne, Australia, 2017: 1725–1731. [7] LIPTON Z C. A critical review of recurrent neural networks for sequence learning[J]. arXiv: 1506.00019, 2015. [8] HOCHREITER S and SCHMIDHUBER J. Long short-term memory[J]. Neural Computation, 1997, 9(8): 1735–1780. doi: 10.1162/neco.1997.9.8.1735 [9] CHO K, VAN MERRIËNBOER B, GULCEHRE C, et al. Learning phrase representations using RNN encoder–decoder for statistical machine translation[C]. The 2014 Conference on Empirical Methods in Natural Language Processing, Doha, Qatar, 2014: 1724–1734. [10] WANG Shoujin, HU Liang, WANG Yan, et al. Sequential recommender systems: Challenges, progress and prospects[C]. The Twenty-Eighth International Joint Conference on Artificial Intelligence, Macao, China, 2019: 6332–6338. [11] MIKOLOV T, CHEN Kai, CORRADO G, et al. Efficient estimation of word representations in vector space[C]. The 1st International Conference on Learning Representations, Scottsdale, USA, 2013. [12] BAEK J W and CHUNG K Y. Multimedia recommendation using Word2Vec-based social relationship mining[J]. Multimedia Tools and Applications, 2021, 80(26): 34499–34515. doi: 10.1007/s11042-019-08607-9 [13] ESMELI R, BADER-EL-DEN M, and ABDULLAHI H. Using Word2Vec recommendation for improved purchase prediction[C]. 2020 International Joint Conference on Neural Networks, Glasgow, UK, 2020: 1–8. [14] 王瑞平, 贾真, 刘畅, 等. 基于DeepFM的深度兴趣因子分解机网络[J]. 计算机科学, 2021, 48(1): 226–232. doi: 10.11896/jsjkx.191200098WANG Ruiping, JIA Zhen, LIU Chang, et al. Deep interest factorization machine network based on DeepFM[J]. Computer Science, 2021, 48(1): 226–232. doi: 10.11896/jsjkx.191200098 [15] 陈一文. 一种改进的基于DeepFM算法的高效CTR预估方法[D]. [硕士论文], 吉林大学, 2020.CHEN Yiwen. An efficient CTR prediction method based on improved DeepFM algorithm[D]. [Master dissertation], Jilin University, 2020. [16] CHEN Qiwei, ZHAO Huan, LI Wei, et al. Behavior sequence transformer for e-commerce recommendation in Alibaba[C]. The 1st International Workshop on Deep Learning Practice for High-Dimensional Sparse Data, Anchorage, Alaska, 2019: 12. [17] MA Jiaqi, ZHAO Zhe, YI Xinyang, et al. Modeling task relationships in multi-task learning with multi-gate mixture-of-experts[C]. The 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, London, United Kingdom, 2018: 1930–1939. [18] ZHU Han, JIN Junqi, TAN Chang, et al. Optimized cost per click in Taobao display advertising[C]. The 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Halifax, Canada, 2017: 2191–2200. -
图(3) / 表(2)
计量
- 文章访问数: 822
- HTML全文浏览量: 851
- PDF下载量: 145
- 被引次数: 0
下载:
