[1]苗 宇,金醒男,杜永萍.基于 Multi-Aspect 的融合网络用户画像生成方法[J].计算机技术与发展,2022,32(08):20-25.[doi:10. 3969 / j. issn. 1673-629X. 2022. 08. 004]
　MIAO Yu,JIN Xing-nan,DU Yong-ping.A User Profile Generation Method Based on Multi-Aspect Converged Network[J].,2022,32(08):20-25.[doi:10. 3969 / j. issn. 1673-629X. 2022. 08. 004]
点击复制

基于 Multi-Aspect 的融合网络用户画像生成方法()
分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
32
期数:
2022年08期
页码:
20-25
栏目:
大数据分析与挖掘
出版日期:
2022-08-10

文章信息/Info

Title:
A User Profile Generation Method Based on Multi-Aspect Converged Network
文章编号:
1673-629X(2022)08-0020-06
作者:
苗 宇; 金醒男; 杜永萍
北京工业大学 信息学部,北京 100124
Author(s):
MIAO Yu; JIN Xing-nan; DU Yong-ping
Faculty of Information Technology,Beijing University of Technology,Beijing 100124,China
关键词:
用户画像; 多层级特征提取; 关键词抽取; 循环神经网络; 注意力机制
Keywords:
user profile; multi-level feature extraction; keyword extraction; recurrent neural network; attention mechanism
分类号:
TP391
DOI:
10. 3969 / j. issn. 1673-629X. 2022. 08. 004
摘要:
在大数据时代,用户画像对于企业了解并获取目标用户的重要性日益提升,但基于统计的用户画像方法无法处理非结构化的文本数据,而传统的基于模型的用户画像方法亦无法从多角度深层次提取用户特征。 为实现更加全面且精准的用户属性预测,该文提出一种基于多层级特征提取的融合网络用户画像生成方法,通过对用户原始文本关键词的提取和排序,分别生成基于 top 2 关键词的子句表示和 top N 关键词的词向量,并结合循环神经网络和注意力机制,构建多层次用户特征提取的分类模型,利用原始用户数据进行用户属性预测。 在搜狗用户搜索文本数据集上的实验表明,文中模型较其他基线模型在分类准确率上显著提升,达到 0. 73,通过消融实验进一步表明各个模块均为有效提取用户特征从而提升分类准确率发挥了重要作用。
Abstract:
In the era of big data,user profile is becoming more and more important for enterprises to understand and obtain target users,but the user profile generation method based on statistics? cannot deal with unstructured text data,and the traditional model - based userprofile generation method cannot deeply extract user characteristics from multiple angles. In order to achieve more comprehensive andaccurate user attribute prediction,we propose a fusion network user profile generation method based on multi-level feature extraction. Byextracting and sorting the user’s original text keywords,the clause representation based on top 2 keywords and the word vector of top Nkeywords are generated respectively. Combined with recurrent neural network and attention mechanism,the classification model of multi-level user feature extraction is constructed,and the user attribute prediction is carried out by using the original user data. The experimenton Sogou user search text dataset shows that the classification accuracy of the proposed model is significantly improved compared withother baseline models,reaching 0. 73. The ablation experiments further show that each module plays an important role in effectivelyextracting user features, so as to improve the classification accuracy.

相似文献/References:

[1]吴明礼,杨双亮. 基于移动特征数据的内容推送技术研究与应用[J].计算机技术与发展,2017,27(09):155.
　WU Ming-li,YANG Shuang-liang. Research and Application on Content Push Technology with Mobile Feature Data[J].,2017,27(08):155.
[2]姜红玉,汪 朋,封 雷.基于流式计算的实时用户画像系统研究[J].计算机技术与发展,2020,30(07):186.[doi:10. 3969 / j. issn. 1673-629X. 2020. 07. 039]
　JIANG Hong-yu,WANG Peng,FENG Lei.Research on Real-time User Profile System Based on Stream Computing[J].,2020,30(08):186.[doi:10. 3969 / j. issn. 1673-629X. 2020. 07. 039]
[3]罗利能,吴秋蓉,石莹禹,等.智能化数据挖掘学习平台的设计与实现[J].计算机技术与发展,2021,31(05):168.[doi:10. 3969 / j. issn. 1673-629X. 2021. 05. 029]
　,,et al.ＤｅｓｉｇｎａｎｄＩｍｐｌｅｍｅｎｔａｔｉｏｎｏｆＩｎｔｅｌｌｉｇｅｎｔＤａｔａＭｉｎｉｎｇＬｅａｒｎｉｎｇＰｌａｔｆｏｒｍ[J].,2021,31(08):168.[doi:10. 3969 / j. issn. 1673-629X. 2021. 05. 029]
[4]谢君臣,李 涛*,黄 甫,等.面向药店会员用户画像的构建及应用研究[J].计算机技术与发展,2022,32(03):145.[doi:10. 3969 / j. issn. 1673-629X. 2022. 03. 024]
　XIE Jun-chen,LI Tao*,HUANG Fu,et al.Research on Construction and Application of User Portrait for Pharmacy Members[J].,2022,32(08):145.[doi:10. 3969 / j. issn. 1673-629X. 2022. 03. 024]
[5]师奥翔,张 洁.基于改进 RFM 模型的电商用户价值分类的研究[J].计算机技术与发展,2022,32(12):123.[doi:10. 3969 / j. issn. 1673-629X. 2022. 12. 019]
　SHI Ao-xiang,ZHANG Jie.Research on E-commerce User Value Classification Based on Improved RFM Model[J].,2022,32(08):123.[doi:10. 3969 / j. issn. 1673-629X. 2022. 12. 019]

更新日期/Last Update: 2022-08-10