Tag relatedness in image folksonomies

Abstract

Folksonomies-networks of users, resources, and tags-allow users to retrieve, organize and browse web contents. However, their advantages are limited mainly due to the noisiness of user provided tags. To overcome this issue, we propose an approach for characterizing related tags in folksonomies: we use tag co-occurrence statistics and Laplacian Score based feature selection in order to create empirical co-occurrence probability distribution for each tag; then we identify related tags on the basis of the dissimilarity between their distributions. To this purpose we introduce variant of the Jensen-Shannon Divergence, which is more robust to statistical noise. We experimentally evaluate our approach by using Word-Net and compare it to a common tag-relatedness approach based on the cosine similarity. The results show the advantage of our approach over the competing method.

Folksonomias e pós-verdade: desafios para a organização do conhecimento

“…Segundo Mousselly-Sergieh et al (2014) a folksonomia é uma estrutura relacional composta por recursos informacionais, tags e usuários que possibilita a organização, navegação e recuperação de conteúdos na web. Nela observa-se a facilidade de acesso aos conteúdos, posto que são sistematizados de modo personalizável, por meio da linguagem natural.…”

K-Means Clustering in Dual Space for Unsupervised Feature Partitioning in Multi-view Learning

“…Consider a corpus of documents, such as a literary corpus or a corpus of web pages (for now we disregard the hyper-links and the multimedia content and focus on text only). Each document, under the bag-of-words representation (that disregards the structure of the text [9]- [13]), can be represented by the count of the occurrences of each word of a reference dictionary. This representation can take the form of a table, where each document corresponds to a row and each word to a column (we call it the [row=document,column=word] representation): each table-cell contains the count of the number of occurrences of a word into a document.…”