摘要
随着互联网技术的快速发展,用户数量日益增长,社交网络平台对周期性智能推荐的需求也日益增加。为了解决当前智能推荐算法准确率低、推荐速度慢等问题,提出了一种基于挖掘算法的用户大数据周期智能推荐算法。算法首先采用协同推荐算法对用户历史行为进行分析,并通过数据相似性衡量智能推荐的效果,使用Top-N算法优化推荐过程,达到周期智能推荐的目的;然后采用基于神经网络的挖掘算法对智能推荐算法进行优化,挖掘长时数据关系的同时保持短时数据之间的非线性;最后通过引入灰色均衡算法对相似度计算优化,从而缩短推荐时间。实验结果表明,所提算法在相似度计算准确度方面提升7%,推荐精确度提升6%,召回率提升8%,有效地提高了数据周期智能推荐的精读和效率,提高了个性化服务的质量。
Big data has become an economic asset,and contains a large amount of information data.Once the server is attacked,massive private information will be leaked.In order to achieve the safe sharing and utilization of big data,this paper proposed a method of monitoring anonymous big data access risk based on deep confrontation learning.Firstly,we analyzed the access risk factors from three aspects:subject,object and environment,including access time,authority,data sensitivity and network delay.Then,we used generator and discriminator to generate a deep confrontation learning network,and took the relevant data of risk factors as the network input to extract the risk characteristics.Moreover,we used the information entropy algorithm to calculate the risk value and set a risk threshold.Finally,we constructed a discriminant function to monitor the risk of anonymous big data access.The experimental results show that the proposed method has strong feature learning ability,avoiding high system throughput in the monitoring process.And the monitoring results are accurate.
作者
陈云云
刘永山
CHEN Yun-yun;LIU Yong-shan(Department of Network Security,Shanxi Police College,Taiyuan Shanxi 030400,China;Department of Information Science and Engineering,Yanshan University,Qinhuangdao Hebei 066044,China)
出处
《计算机仿真》
2024年第4期456-460,465,共6页
Computer Simulation
基金
山西省哲学规划办公室(2022YD165)。
关键词
数据挖掘算法
用户大数据
推荐算法
卷积神经网络
Deep confrontation learning
Anonymous big data
Access risk monitoring
Information entropy algorithm
Discriminant function
