摘要
对云环境下多载体图书信息进行分类能够提高多载体图书信息的利用效率。针对当前多载体图书信息分类方法存在的分类准确度较低,分类速度较慢问题,提出一种基于LDA模型的云环境下多载体图书信息自动分类方法,根据多载体图书信息在影射空间中的Euclidean距离,确定多载体图书信息特征聚类的目标函数,根据目标函数初始化聚类中心,实现核函数影射构造,确定图书信息样本到聚类中心的距离,根据到聚类中心的距离,对核矩阵进行改进,构建多载体图书信息特征提取模型。利用构建的模型,通过吉布斯参数采样,构建LDA模型,根据LDA模型中图书信息主题分布和特征词分布的后验概率,完成多载体图书信息自动分类。实验结果表明,所提方法进行图书信息分类,分类速度较快,且分类的准确度较高。
This article puts forward a method to automatically classify multi-carrier book information in cloud environment based on LDA model.According to Euclidean distance of multi-carrier book information in mapping space,we could determine the objective function of feature clustering of multi-carrier book,and then initialized the clustering center based on objective function to realize the Kernel function mapping.Moreover,we determined the distance from the book information sample to the cluster center.Based on this distance,we improved the kernel matrix and built the model to extract features of multi-carrier book information.In addition,we used the constructed model and Gibbs parameter sampling to build LDA model.According to the posteriori probability of topic distribution and distribution of characteristic words of book information in LDA model,we could complete the automatic classification of multi-carrier book information.Simulation results prove that the proposed method has high classification speed and high classification accuracy.
作者
廖辰刚
LIAO Chen-gang(Sichuan Normal University,library,Chengdu,Sichuan 610068,China)
出处
《计算机仿真》
北大核心
2019年第4期358-361,368,共5页
Computer Simulation
关键词
云环境
多载体
图书信息
自动分类
Cloud environment
Multi-carrier
Book information
Automatic classification
