the topic-perspective model for social tagging systems
DESCRIPTION
The Topic-Perspective Model for Social Tagging Systems. 蔡跳. INTRODUCTION. social data--social annotations--tags a new type of information source tag recommendation、prediction 、clustering、classification、IR. Tags. RELATED WORK1. Topic Analysis using Generative Models text mining: - PowerPoint PPT PresentationTRANSCRIPT
The Topic-Perspective Model for Social Tagging Systems
蔡跳
INTRODUCTION
• social data--social annotations--tags
• a new type of information source
• tag recommendation、 prediction 、 clustering、 classification、IR
Tags
RELATED WORK1
• Topic Analysis using Generative Models
text mining:
1.Naïve Bayesian model,
2.Probabilistic Latent Semantic Indexing (PLSI) model,
3.Latent Dirichlet Allocation (LDA) model• correlated LDA, switchLDA, Link-LDA, To
pic-Link LDA
RELATED WORK2
• Generative Models for Social Tagging
1.Conditionally-independent LDA (CI-LDA) model
2. Community-based categorical annotation (CCA) model
3.correlated or correspondence LDA (CorrLDA) model
DXK doc-topic分布)(w KXW topic-word分布)(t KXT topic-tag分布
Topic-Perspective Model
• 真实模拟 annotation的生成过程, user 、 document、 word、 tag统一在一个模型中
• motivation:表示和连接可见的及不可见的变量
• Output: user perspective可用于个性化搜素
)(d DXK doc-topic分布)(w KXW topic-word分布)(t KXT topic-tag分布
LXT persp-tag分布
)(u UXL user-persp分布
a vector indicating the probability each tag is generated from topics
Parameter Estimation
• Variational expectation maximization
• Expectation propagation
• Gibbs sampling
Parameter Estimation
Parameter Estimation
Experiments and results
• Datasets: del.icio.us, 1-2 2009, 41190 documents, 4414 users, 28740 tags, 129908 words, 10% test, 90% train
• Evaluation Criterion: perplexity.概括归纳新文档的 tags的能力
Experiment Setup
• Topic K Perspective L 的选择
Results
Discovered topics and perspectives
• 谢谢