|
| 基于LDA模型的WEB文本分类 |
| Web Text Classification based on LDA Model |
| |
| DOI：10.3969/j.issn.1671-5322.2009.04.016 |
| 中文关键词: LDA 主题模型 WEB分类 |
| 英文关键词: Latent Dirichlet Allocation(LDA) topic model WEB classification |
| 基金项目: |
|
| 摘要点击次数: 5963 |
| 全文下载次数: 6085 |
| 中文摘要: |
| 提出了基于LDA(Latent Dirichlet Allocation)主题模型的Web文本分类方法,利用MCMC方法中的Gibbs抽样获得模型参数从而获取词汇的概率分布,使隐藏于WEB文本内的不同主题与WEB文本字词建立关系.将LDA算法应用于WEB文本分类识别领域,在实验中与k均值聚类和贝叶斯网络方法进行了对比,其结果表明LDA与其他同类算法相比具有一定的优势. |
| 英文摘要: |
| A kind of web text classification is put forward on the basis of LDA model.Latent Dirichlet Allocation(LDA) is an unsupervised topic learning model which extracts latent topics from text data.Parameters are estimated with Gibbs sampling of MCMC and the word probability is represented.Thus different latent topics are associated with observable words.Contrasting to SVM and Bayesian Network,the result in the experiment shows that LDA has the better performance than any other algorithm. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |