Clustering as an approach to support the automatic definition of semantic hyperlinks

Camacho-Guerrero, José Antonio; Carvalho, Alex Assis de; Pimentel, Maria G.; Munson, Ethan V.; Macedo, Alessandra Alaniz

doi:10.1145/1286240.1286264

Cited by 5 publications

(4 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…With regard to the methodologies for automated linking of textual content, many existing approaches are rather straightforward. For example, approaches to hyperlinking are often restricted to clustering of content [8] or linking via a shared terminology or common named entities [7], which in our case are not useful as the e-commerce consumers use a quite different language than the sellers. To deal with vocabulary differences, several approaches have been proposed, including text normalization, query reformulation, search results clustering, and automatic query expansion.…”

Section: Related Workmentioning

confidence: 99%

“…The vocabulary size (number of unique words) of the full training collection is 102,077. Out of these, 8% (8,626) are found only in the product descriptions and not in reviews; 65% (65,898) are found only in the reviews (not in products); and 27% (27,553) are shared between the two. We see that users tend to employ many words already mentioned in the descriptions, but they also add a large set of new words not previously used to describe the items.…”

Section: Training Collectionmentioning

confidence: 99%

“…The annotators were presented both the text and image of the pin. Additionally, the annotators were presented with all the webshops, containing both the textual and visual information from all the products 8 . For each query, the annotator indicated the path to all the relevant webshops.…”

Section: Queries and Annotationsmentioning

confidence: 99%

See 2 more Smart Citations

Latent Dirichlet allocation for linking user-generated content and e-commerce data

Zoghbi

Vulić

Moens

2016

Information Sciences

View full text Add to dashboard Cite

Automatic linking of online content improves navigation possibilities for end users. We focus on linking content generated by users to other relevant sites. In particular, we study the problem of linking information between different usages of the same language, e.g., colloquial and formal idioms or the language of consumers versus the language of sellers. The challenge is that the same items are described using very distinct vocabularies. As a case study, we investigate a new task of linking textual Pinterest.com pins (colloquial) to online webshops (formal). Given this task, our key insight is that we can learn associations between formal and informal language by utilizing aligned data and probabilistic modeling. Specifically, we thoroughly evaluate three different modeling paradigms based on probabilistic topic modeling: monolingual latent Dirichlet allocation (LDA), bilingual LDA (BiLDA) and a novel multi-idiomatic LDA model (MiLDA). We compare these to the unigram model with Dirichlet prior. Our results for all three topic models reveal the usefulness of modeling the hidden thematic structure of the data through topics, as opposed to the linking model based solely on the standard unigram. Moreover, our proposed MiLDA model is able to deal with intrinsic multiidiomatic data by considering the shared vocabulary between the aligned document pairs. The proposed MiLDA obtains the largest stability (less variation with changes in parameters) and highest mean average precision scores in the linking task.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Training Collectionmentioning

confidence: 99%

See 1 more Smart Citation

Latent Dirichlet allocation for linking user-generated content and e-commerce data

Zoghbi

Vulić

Moens

2016

Information Sciences

View full text Add to dashboard Cite

show abstract

“…With respect to works dealing only with semantic hyperlinks and automatic hyperlinking, on one hand, [14] presents an approach that intends to create semantic links through the use of clustering (which is done with the Wordnet English lexicon) and Latent Semantic Linking. On the other hand, [29], does automatic linking of plain text Chinese news by utilizing Self-Organized Maps in order to create groups of words and groups of documents that allow to establish different relationships among the existing documents of the corpus; intra-cluster, inter-cluster, and aggregate links are finally constructed.…”

Section: Related Workmentioning

confidence: 99%

Topic Calculation and Clustering: An Application to Wikipedia

Garza

Brena

Ramírez

2008

2008 Seventh Mexican International Conference on Artificial Intelligence

View full text Add to dashboard Cite

Wikipedia is nowadays one of the most valuable information resources; nevertheless, its current structure, which has no formal organization, does not allow to always have a useful browsing among topics. Moreover, even though most Wikipedia pages include a "See Also" section for navigating through those articles' related Wikipedia pages, the only references included here are those which authors are aware of, leading to incompleteness and other irregularities. In this work, a method for finding related Wikipedia articles is proposed; this method relies on a framework that clusters documents into semantically-calculated topics and selects the closest documents which could enrich the "See Also" section.

show abstract

Feedback-driven clustering for automated linking of web pages

Oest

Rege

2013

8th International Conference for Internet Technology and Secured Transactions (ICITST-2013)

View full text Add to dashboard Cite

Clustering as an approach to support the automatic definition of semantic hyperlinks

Cited by 5 publications

References 10 publications

Latent Dirichlet allocation for linking user-generated content and e-commerce data

Latent Dirichlet allocation for linking user-generated content and e-commerce data

Topic Calculation and Clustering: An Application to Wikipedia

Feedback-driven clustering for automated linking of web pages

Contact Info

Product

Resources

About