一种基于高层特征融合的网络商品分类

doi:10.13190/j.jbupt.2020-033

北京邮电大学学报 ›› 2020, Vol. 43 ›› Issue (5): 98-104,117.doi: 10.13190/j.jbupt.2020-033

一种基于高层特征融合的网络商品分类

刘逸琛, 孙华志, 马春梅, 姜丽芬, 钟长鸿

天津师范大学计算机与信息工程学院, 天津 300387

收稿日期:2020-04-23 发布日期:2021-03-11
通讯作者: 马春梅(1985-),女,讲师,E-mail:mcmxhd@163.com. E-mail:mcmxhd@163.com
作者简介:刘逸琛(1995-),男,硕士生.
基金资助:
国家自然科学基金项目（61702370）；天津市自然科学基金项目（18JCYBJC85900，18JCQNJC70200）；天津市科技发展战略研究计划项目（17ZLZXZF00530）；天津市教委科研计划项目（JW1702）

Commodity Classification of Online Based on High-Level Feature Fusion

LIU Yi-chen, SUN Hua-zhi, MA Chun-mei, JIANG Li-fen, ZHONG Chang-hong

School of Computer and Information Engineering, Tianjin Normal University, Tianjin 300387, China

Received:2020-04-23 Published:2021-03-11

摘要/Abstract

摘要： 为了利用商品文本标题实现商品自动分类，提出一种基于高层特征融合的商品分类模型.首先，提出基于字嵌入和词嵌入的文本底层特征表示法，进而获得更强的商品标题结构特征表达；其次，提出了联合自注意力、卷积神经网络和通道注意力的机制，对文本标题的底层特征进行增强并获得高层增强特征；最后，通过将文本的字嵌入和词嵌入的高层增强特征进行融合，最终获得商品文本标题的综合特征，并实现商品自动分类.以商品标题语料作为数据集进行了实验，实验结果表明，该模型对三级商品类别的分类精度能够达到84.348%，召回率和F1值分别达到了47.8%和49.4%，优于现有可用于商品文本标题分类的先进短文本分类方法.

关键词: 商品分类, 短文本分类, 特征融合, 特征增强, 注意力机制

Abstract: In order to realize automatic classification of commodities by leveraging text titles of commodities, a commodity classification model high-level feature fusion (HFF) based on high-level feature fusion is proposed. Firstly, a char embedding and word embedding based low-level feature representation method for the text title is proposed. Then a stronger feature expression of the commodity title structure can be obtained. Secondly, a joint self-attention mechanism, convolutional neural network, and channel attention are proposed to enhance the low-level features and obtain high-level enhancement features of the text title. Finally,by fusing the high-level enhancement features of the word embedding and the char embedding of the text, a comprehensive feature of the text title of the commodity is finally obtained and used for the commodity classification. Experiments are conduct on the dataset of the commodity titles. The experiments show that the classification accuracy of HFF for the third-level commodity can reach 84.348%. In addition, the recall and the F1 value of the HFF reach 47.8% and 49.4%, respectively, which is superior to the existing advanced short text classification method that can be used for the commodity text titles classification.

Key words: commodity classification, short text classification, feature fusion, feature enhancement, attention

中图分类号:

TP183

刘逸琛, 孙华志, 马春梅, 姜丽芬, 钟长鸿. 一种基于高层特征融合的网络商品分类[J]. 北京邮电大学学报, 2020, 43(5): 98-104,117.

LIU Yi-chen, SUN Hua-zhi, MA Chun-mei, JIANG Li-fen, ZHONG Chang-hong. Commodity Classification of Online Based on High-Level Feature Fusion[J]. Journal of Beijing University of Posts and Telecommunications, 2020, 43(5): 98-104,117.

参考文献

[1] Xu Jingyun, Cai Yi. Incorporating context-relevant knowledge into convolutional neural networks for short text classification[C]//Proc of the 33rd AAAI Conf on Artificial Intelligence. Honolulu:AAAI, 2019:10067-10068.
[2] Ma Chenglong, Xu Weiqun, Li Peijia, et al. Distributional representations of words for short text classification[C]//Proc of the 51nd Conf of the North American Chapter of the Association for Computational Linguistics:Human Language Technologies(HLT-NAACL). Stroudsburg:ACL, 2015:33-38.
[3] Danesh I, Steve W, Calton P. Study of static classification of social spam profiles in myspace[C]//Proceedings of the 4th Int AAAI Conf on Weblogs and Social Media(ICWSM). Menlo Park:AAAI, 2010:70-75.
[4] Joao P A V, Raimundo S M. An analysis of convolutional neural networks for sentence classification[C]//Proceedings of Conferencia Latinoamericana En Informatica(CLEI). Piscataway:IEEE, 2017:1-5.
[5] Xie Jun, Chen Bo, Gu Xinglong, et al. Self-attention-based bilstm model for short text fine-grained sentiment classification[J]. IEEE Access, 2019, 7:180558-180570.
[6] 朱皓, 谭咏梅. 基于胶囊的英文文本蕴含识别方法[J]. 北京邮电大学学报, 2019, 42(3):21-28. Zhu Hao, Tan Yongmei. English textual entailment recognition using capsules[J]. Journal of Beijing University of Posts and Telecommunications, 2019, 42(3):21-28.
[7] Ashish V, Noam S, Niki P, et al. Attention is all you need[C]//Proc of the 30th Advances in Neural Information Processing Systems(NIPS). Cambridge:MIT Press, 2017:6000-6010.
[8] Luo Ruixuan, Xu Jingjing, Zhang Yi, et al. PKUSEG:a toolkit for multi-domain Chinese word segmentation[EB/OL]. 2019(2019-06-27)[2020-07-18]. https://arxiv.org/abs/1906.11455.
[9] Tomas M, Kai C, Greg C, et al. Efficient estimation of word representations in vector space[C]//Proc of the 1rd Int Conf on Learning Representations(ICLR). Scottsdale:ICLR, 2013:1-12.
[10] Reuven Y. Rubinstein. Optimization of computer simulation models with rare events[J]. European Journal of Operational Research, 1997, 99(1):89-112.
[11] GitHub. DFF-Dataset[EB/OL]. 2020(2020-06-18)[2020-10-04].https://github.com/liuyichenaa1/DFF-Dataset.
[12] Adam P, Sam G, Francisco M, et al. pyTorch:an imperative style, high-performance deep learning library[C]//Proc of the 32th Advances in Neural Information Processing Systems(NeurIPS 2019). Vancouver:MIT Press, 2019:8024-8035.
[13] Guo Long, Zhang Dongxiang, Wang Lei, et al. CRAN:a hybrid CNN-RNN attention-based model for text classification[C]//Proc of the 37th International Conference on Conceptual Modeling(ER2018). Berlin:Springer-Verlag, 2018:571-585.
[14] Jiang Wei, Jin Zhong. Integrating bidirectional LSTM with inception for text classification[C]//Proc of the 4th IAPR Asian Conference on Pattern Recognition(ACPR 2017). Piscataway:IEEE, 2017:870-875.

[1]	包乾辉文娟石淑珍董萌萍刘雪. 融合句法依存与 BERT-Att-BiLSTM 的属性分类[J]. 北京邮电大学学报, 2023, 46(4): 123-128.
[2]	周雨昊孙哲吴晓非禹可. 基于门控特征融合的中文错别字纠正模型[J]. 北京邮电大学学报, 2023, 46(4): 91-96,122.
[3]	苏兴禹可吴晓非. 基于层次门控交互融合网络的谣言检测方法[J]. 北京邮电大学学报, 2023, 46(4): 97-102.
[4]	胡敏熊思黄宏程张光华王春芳. 基于RAPNet的脑肿瘤MRI图像三维分割[J]. 北京邮电大学学报, 2023, 46(2): 91-97.
[5]	宋宇博高嘉振. 改进YOLOv3算法的交通多目标检测方法[J]. 北京邮电大学学报, 2022, 45(5): 103-108.
[6]	李雪, 李晓艳, 王鹏, 孙梦宇, 吕志刚. 结合双注意力与特征融合的孪生网络目标跟踪[J]. 北京邮电大学学报, 2022, 45(4): 116-122.
[7]	刘琳岚, 宋修洋, 陈宇斌. 基于网络表示学习的机会网络链路预测[J]. 北京邮电大学学报, 2022, 45(4): 64-69,103.
[8]	晁浩, 马庆敏, 刘永利. 利用卷积回声状态网络实现脑电情感识别[J]. 北京邮电大学学报, 2022, 45(2): 36-43.
[9]	王科平, 韦金阳, 杨艺, 费树岷, 崔科飞. 融合特征增强及多尺度损失的增量去雾算法[J]. 北京邮电大学学报, 2022, 45(2): 57-64.
[10]	宁静, 佘红艳, 赵东, 罗丹, 王磊. 一种路网级交通事故风险预测方法[J]. 北京邮电大学学报, 2022, 45(2): 72-78.
[11]	赵海英, 朱会, 侯小刚. 基于改进EMA单元的传统服饰图像语义分割[J]. 北京邮电大学学报, 2022, 45(1): 69-74.
[12]	李威, 陈曙东, 欧阳小叶, 杜蓉, 王荣. 基于高速多核网络的远监督关系抽取方法[J]. 北京邮电大学学报, 2020, 43(5): 71-76.
[13]	王小捷, 白子薇, 李可, 袁彩霞. 机器阅读理解的研究进展[J]. 北京邮电大学学报, 2019, 42(6): 1-9.
[14]	朱皓, 谭咏梅. 基于胶囊的英文文本蕴含识别方法[J]. 北京邮电大学学报, 2019, 42(3): 21-28.
[15]	吕光瑞, 蔡国永, 林煜明. 基于多模态判别性嵌入空间的图像情感分析[J]. 北京邮电大学学报, 2019, 42(1): 61-67.

一种基于高层特征融合的网络商品分类

Commodity Classification of Online Based on High-Level Feature Fusion

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

本文评价