摘要
为提升网络个性化服务质量,针对传统推荐算法计算结果精度不高、覆盖率不全面的问题,提出一种自注意力机制下网络用户行为数据推荐方法。引入自注意力机制,模拟人类大脑处理看到事物时独有的信号处理机制,通过用户潜在偏好的集合与行为特征集合构建用户行为偏好模型,挖掘用户网络行为特点;将网络用户行为拟作个体评分问题,把用户评分信息当作区间型符号数据,在改进Hausdorff距离方法的前提下,使用K均值聚类算法划分区间型符号数据,预测目标用户对网络行为的评分,利用最近邻评分原则选取评分最高的项目作为最优推荐信息推送。以真实数据集作为仿真样本,实验结果证明所提方法的数据推荐准确率高、推荐内容覆盖范围广、实用性强,可广泛应用于各大门户网站。
In order to improve the network personalized service, this paper presented a method of recommending network user behavior data based on self-attention mechanism. Firstly, self-attention mechanism was introduced to simulate the unique signal processing mechanism of human beings when brains processed what they saw. And then, a model of user behavior preference was constructed through the set of user potential preferences and behavior characteristics, for mining the characteristics of user network behaviors. On this basis, the network user behavior can be considered as an individual scoring problem. Meanwhile, the user scoring information was regarded as the interval symbolic data. On the premise of improving the Hausdorff distance method, K-means clustering algorithm was used to divide the interval symbolic data, and thus to predict the target user’s scores on network behaviors. Finally, the nearest neighbor scoring principle was used to select the item with the highest score as the optimal recommendation information. The real dataset was taken as a simulation sample. Experimental results show that the proposed method has high accuracy of data recommendation wider recommendation content and stronger practicability, so it can be widely applied in major portal websites.
作者
王冲
赵艺璇
汪子尧
WANG Chong;ZHAO Yi-xuan;WANG Zi-yao(College of Computer and Information Security,Guilin University of Electronic Technology,Guilin 541004,Guangxi,China)
出处
《计算机仿真》
北大核心
2022年第12期497-501,共5页
Computer Simulation
基金
国家自然科学基金(72061008)
广西区自然科学基金(2018GXNSFAA294123)
广西可信软件重点实验室基金(kx201923)
2019年研究生创新项目(C21YJM00SJ0S)
2020年研究生创新项目(C21YJM00SJ0H,C21YJM00SJ07)。
关键词
自注意力机制
用户行为
数据推荐
行为偏好
区间型符号
Self-attention mechanism
User behavior
Data recommendation
Behavioral preference
Interval symbol
作者简介
王冲(1972-),男(汉族),四川宣汉人,硕士,教授,主要研究方向为大数据技术,教学信息化,信息资源管理。;赵艺璇(1998-),女(汉族),广东中山人,硕士研究生在读,主要研究方向为推荐算法,自然语言处理。;汪子尧(1998-),男(汉族),四川绵阳人,硕士研究生在读,主要研究方向为推荐算法,信息检索技术。
