Proceedings of the 29th Annual International ACM SIGIR Conference on Research and Development in Information Retrieval 2006
DOI: 10.1145/1148170.1148204
|View full text |Cite
|
Sign up to set email alerts
|

LDA-based document models for ad-hoc retrieval

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

6
591
4
5

Year Published

2008
2008
2014
2014

Publication Types

Select...
4
2
2

Relationship

0
8

Authors

Journals

citations
Cited by 825 publications
(606 citation statements)
references
References 15 publications
6
591
4
5
Order By: Relevance
“…It has been shown to perform as well or better than many other popular techniques for machine learning, data mining, and supervised and unsupervised classification of data. Indeed, LDA has been found to have a similar running time for processing as k-means (Wei and Croft 2006), a long-used approach for unsupervised clustering, which lacks LDA's capability to associate documents with a distribution over topics rather than assignment of each document to a single, unique topic. Modifications, extensions, improvements, and additions to LDA are being developed and released at a rapid pace; some relevant extensions are discussed later in this article.…”
Section: Latent Dirichlet Allocationmentioning
confidence: 99%
“…It has been shown to perform as well or better than many other popular techniques for machine learning, data mining, and supervised and unsupervised classification of data. Indeed, LDA has been found to have a similar running time for processing as k-means (Wei and Croft 2006), a long-used approach for unsupervised clustering, which lacks LDA's capability to associate documents with a distribution over topics rather than assignment of each document to a single, unique topic. Modifications, extensions, improvements, and additions to LDA are being developed and released at a rapid pace; some relevant extensions are discussed later in this article.…”
Section: Latent Dirichlet Allocationmentioning
confidence: 99%
“…The explicit assumption about the prior sources of these variables provides complete generative semantics for the model [2][6] [16]. Moreover, the mathematical property that the Dirichlet priors of p(θ d | α) and p(Ф z | β) are conjugate to their likelihoods (multinomial distributions) p(z| θ d ) and p(w| Ф z ) results in the fact that their posteriors p(θ d | α, {z i | for all tokens in doc d}) and p(Ф z | β, {w i | for all tokens generated by z}) are also Dirichlet distributions.…”
Section: Fig 1 Latent Dirichlet Relevance Modelmentioning
confidence: 99%
“…In LDA, topic proportion of every document is a K-dimensional hidden variable randomly drawn from the same Dirichlet distribution, where K is the number of topics. Thus, generative semantics of LDA are complete [16]. LDA and its variants have been applied in many applications such as finding scientific topics [6], E-community discovery [18], mixedmembership analysis [5] and ad-hoc retrieval for representing document language model [4] [16].…”
Section: Related Workmentioning
confidence: 99%
See 1 more Smart Citation
“…LDA is an intensively studied model, and the experiments are really impressive compared to other known information retrieval techniques. The applications of LDA include entity resolution [4], fraud detection in telecommunication systems [5], image processing [6,7,8] and ad-hoc retrieval [9].…”
Section: Introductionmentioning
confidence: 99%