The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
À medida que o tamanho dos dados dos problemas de classificação multi-rótulos relacionados à Web continua a aumentar, o espaço de rótulos também cresceu extremamente. Por exemplo, o número de rótulos que aparecem em tarefas de marcação de páginas da Web e de recomendação de comércio eletrônico chega a centenas de milhares ou até milhões. Neste artigo, propomos uma árvore de particionamento de grafos (GPT), que é uma nova abordagem para aprendizagem multi-rótulo extrema. Em um nó interno da árvore, o GPT aprende um separador linear para particionar um espaço de recursos, considerando aproximado k-gráfico do vizinho mais próximo dos vetores de rótulo. Também desenvolvemos um procedimento simples de otimização sequencial para aprender os classificadores binários lineares. Extensos experimentos em conjuntos de dados do mundo real em grande escala mostraram que nosso método alcança melhor precisão de previsão do que métodos baseados em árvore de última geração, ao mesmo tempo que mantém uma previsão rápida.
Yukihiro TAGAMI
Yahoo Japan Corporation,Kyoto University
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copiar
Yukihiro TAGAMI, "Recursive Nearest Neighbor Graph Partitioning for Extreme Multi-Label Learning" in IEICE TRANSACTIONS on Information,
vol. E102-D, no. 3, pp. 579-587, March 2019, doi: 10.1587/transinf.2018EDP7106.
Abstract: As the data size of Web-related multi-label classification problems continues to increase, the label space has also grown extremely large. For example, the number of labels appearing in Web page tagging and E-commerce recommendation tasks reaches hundreds of thousands or even millions. In this paper, we propose a graph partitioning tree (GPT), which is a novel approach for extreme multi-label learning. At an internal node of the tree, the GPT learns a linear separator to partition a feature space, considering approximate k-nearest neighbor graph of the label vectors. We also developed a simple sequential optimization procedure for learning the linear binary classifiers. Extensive experiments on large-scale real-world data sets showed that our method achieves better prediction accuracy than state-of-the-art tree-based methods, while maintaining fast prediction.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2018EDP7106/_p
Copiar
@ARTICLE{e102-d_3_579,
author={Yukihiro TAGAMI, },
journal={IEICE TRANSACTIONS on Information},
title={Recursive Nearest Neighbor Graph Partitioning for Extreme Multi-Label Learning},
year={2019},
volume={E102-D},
number={3},
pages={579-587},
abstract={As the data size of Web-related multi-label classification problems continues to increase, the label space has also grown extremely large. For example, the number of labels appearing in Web page tagging and E-commerce recommendation tasks reaches hundreds of thousands or even millions. In this paper, we propose a graph partitioning tree (GPT), which is a novel approach for extreme multi-label learning. At an internal node of the tree, the GPT learns a linear separator to partition a feature space, considering approximate k-nearest neighbor graph of the label vectors. We also developed a simple sequential optimization procedure for learning the linear binary classifiers. Extensive experiments on large-scale real-world data sets showed that our method achieves better prediction accuracy than state-of-the-art tree-based methods, while maintaining fast prediction.},
keywords={},
doi={10.1587/transinf.2018EDP7106},
ISSN={1745-1361},
month={March},}
Copiar
TY - JOUR
TI - Recursive Nearest Neighbor Graph Partitioning for Extreme Multi-Label Learning
T2 - IEICE TRANSACTIONS on Information
SP - 579
EP - 587
AU - Yukihiro TAGAMI
PY - 2019
DO - 10.1587/transinf.2018EDP7106
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E102-D
IS - 3
JA - IEICE TRANSACTIONS on Information
Y1 - March 2019
AB - As the data size of Web-related multi-label classification problems continues to increase, the label space has also grown extremely large. For example, the number of labels appearing in Web page tagging and E-commerce recommendation tasks reaches hundreds of thousands or even millions. In this paper, we propose a graph partitioning tree (GPT), which is a novel approach for extreme multi-label learning. At an internal node of the tree, the GPT learns a linear separator to partition a feature space, considering approximate k-nearest neighbor graph of the label vectors. We also developed a simple sequential optimization procedure for learning the linear binary classifiers. Extensive experiments on large-scale real-world data sets showed that our method achieves better prediction accuracy than state-of-the-art tree-based methods, while maintaining fast prediction.
ER -