Abstract | ||
---|---|---|
Blog mining is an important area of behavior informatics because produces effective techniques for analyzing and understanding human behaviors from social media. In this paper, we propose the blogger-link-topic model for blog mining based on the multiple attributes of blog content, bloggers, and links. In addition, we present a unique blog classification framework that computes the normalized document-topic matrix, which is applied our model to retrieve the classification results. After comparing the results for blog classification on real-world blog data, we find that our blogger-link-topic model outperforms the other techniques in terms of overall precision and recall. This demonstrates that additional information contained in blog-specific attributes can help improve blog classification and retrieval results. |
Year | DOI | Venue |
---|---|---|
2011 | 10.1007/978-3-642-28320-8_3 | PAKDD Workshops |
Keywords | Field | DocType |
blogger-link-topic model,blog content,blog mining,additional information,effective technique,blog-specific attribute,classification result,unique blog classification framework,real-world blog data,blog classification,latent dirichlet allocation,classification | Data mining,Informatics,Latent Dirichlet allocation,Social media,Normalization (statistics),Information retrieval,Computer science,Precision and recall,Human behavior,Topic model | Conference |
Citations | PageRank | References |
1 | 0.36 | 12 |
Authors | ||
1 |
Name | Order | Citations | PageRank |
---|---|---|---|
Flora S. Tsai | 1 | 352 | 23.96 |