BoosTexter: A Boosting-based Systemfor Text Categorization
Machine Learning - Special issue on information retrieval
Object Recognition as Machine Translation: Learning a Lexicon for a Fixed Image Vocabulary
ECCV '02 Proceedings of the 7th European Conference on Computer Vision-Part IV
Knowledge Discovery in Multi-label Phenotype Data
PKDD '01 Proceedings of the 5th European Conference on Principles of Data Mining and Knowledge Discovery
Machine Learning
Multilabel Neural Networks with Applications to Functional Genomics and Text Categorization
IEEE Transactions on Knowledge and Data Engineering
The challenge problem for automated detection of 101 semantic concepts in multimedia
MULTIMEDIA '06 Proceedings of the 14th annual ACM international conference on Multimedia
ML-KNN: A lazy learning approach to multi-label learning
Pattern Recognition
Study on Support Vector Machine Based Decision Tree and Application
FSKD '08 Proceedings of the 2008 Fifth International Conference on Fuzzy Systems and Knowledge Discovery - Volume 05
Multi-label Classification Using Ensembles of Pruned Sets
ICDM '08 Proceedings of the 2008 Eighth IEEE International Conference on Data Mining
Boosting Support Vector Machines Successfully
MCS '09 Proceedings of the 8th International Workshop on Multiple Classifier Systems
The WEKA data mining software: an update
ACM SIGKDD Explorations Newsletter
A hybrid SVM based decision tree
Pattern Recognition
Hi-index | 0.00 |
Multi-label classification (MLC) problems abound in many areas, including text categorization, protein function classification, and semantic annotation of multimedia. Issues that severely limit the applicability of many current machine learning approaches to MLC are the large-scale problem and the high dimensionality of the label space, which have a strong impact on the computational complexity of learning. These problems are especially pronounced for approaches that transform MLC problems into a set of binary classification problems for which SVMs are used. On the other hand, the most efficient approaches to MLC, based on decision trees, have clearly lower predictive performance. We propose a hybrid decision tree architecture that utilizes local SVMs for efficient multi-label classification. We build decision trees for MLC, where the leaves do not give multi-label predictions directly, but rather contain SVM-based classifiers giving multi-label predictions. A binary relevance architecture is employed in each leaf, where a binary SVM classifier is built for each of the labels relevant to that particular leaf. We use several real-world datasets to evaluate the proposed method and its competition. Our hybrid approach on almost every classification problem outperforms the predictive performances of SVM-based approaches while its computational efficiency is significantly improved as a result of the integrated decision tree.