摘要
针对云平台图书数据"大集成"下的信息迷航性,为实现投其所好的个性化推送服务,将云计算环境下图书资源的聚合为框架支撑,首先,基于读者阅读浏览行为建构读者兴趣度模型,并融合时间因素和当下时间窗口图书流行度建构评分权重计算模型,以挖掘读者兴趣度的动态变化性;然后,利用LAD主体模型进行兴趣度主题分类挖掘,引入Markov模型,基于转移矩阵测算不同时段读者兴趣度主题及类型的动态变化特征,并据此,进行读者群组的划分,利用基于内容与协同过滤的算法,为读者个体或群体进行自主的图书数据推送服务。最后,通过实验测试,采用上述推送算法获得的读者群组划分误差值最小,MAEij仅为0.18%,且通过推送准确性基本均在0.9以上,推送耗时均值为5 s,各评价指标均存在显著优势,可见其达到了高精准的设计要求。
In view of the information lost under the "big integration" of book data on cloud platform, in order to realize the personalized push service that it likes, will aggregate the book resources in the cloud computing environment as a framework support. First, based on the readers′ reading and browsing behavior, will construct the reader interest model, and integrate the time factor and the current time window book popularity to construct the scoring weight calculation model, so as to mine the dynamic changes of readers′ interest. Then, the LAD main model is used to mine the interest topic classification, Markov model is introduced to measure the dynamic change characteristics of the topics and types of readers′ interest degree in different periods based on the transfer matrix. Based on this, the reader groups are divided, and the algorithm based on content and collaborative filtering is used to provide independent book data push service for individual or group of readers. Finally, through the experimental test, the error value of the reader group partition obtained by the above-mentioned push algorithm is the smallest, MAEij is only 0.18%, and the accuracy of the push is basically above 0.9, and the average push time is 5 s. Each evaluation index has significant advantages. It can be seen that it meets the design requirements of high precision.
作者
杨雯雯
徐小玲
Yang Wenwen;Xu Xiaoling(Public Teaching Department of Yan′an Vocational and Technical College,Yan′an 716000,China;Xi′an Innovation College of Yan′an University,Xi′an 710000,China)
出处
《电子测量技术》
2020年第22期72-76,共5页
Electronic Measurement Technology
基金
陕西省教育科学规划课题:基于大数据的教学质量评价与分析研究(SGH18H466)项目资助。
关键词
兴趣度主题
时变特征
读者群组
资源聚合
interest topic
time varying feature
reader group
resource aggregation
