[PDF][PDF] The value of unlabeled data for classification problems

T Zhang, F Oles - … of the Seventeenth International Conference on …, 2000 - Citeseer
… rest 9583 training data points as unlabeled data to train a transductive version of SVM. The
… the transductive SVM evaluated on the unlabeled data is worse than that of the supervised …

Learning classification with both labeled and unlabeled data

JN Vittaut, MR Amini, P Gallinari - … Helsinki, Finland, August 19–23, 2002 …, 2002 - Springer
… We have performed experiments using neural networks and support vector machines but
for the classification tasks considered here, they did not show any improvement over a simple …

Unlabeled data classification via support vector machines and k-means clustering

L Maokuan, C Yusheng… - … Conference on Computer …, 2004 - ieeexplore.ieee.org
… based on support vector machines to classify unlabeled data more accurately after training
the data labelled by k-means clustering algorithms. For classifying unlabeled data, which is …

[PDF][PDF] Is unlabeled data suitable for multiclass SVM-based web page classification?

A Zubiaga, V Fresno, R Martínez - Proceedings of the NAACL HLT …, 2009 - aclanthology.org
… of unlabeled data for multiclass SVM classification tasks. We … semisupervised SVM approach
to conduct this kind of tasks. … explain how SVM classifiers work for binary classifications, …

A bagging SVM to learn from positive and unlabeled examples

F Mordelet, JP Vert - Pattern Recognition Letters, 2014 - Elsevier
identifying a data of interest is difficult or expensive, eg, because human intervention is
necessary or expensive experiments are needed, while unlabeled dataUnlabeled data are data

Incremental support vector machine for unlabeled data classification

JH Hong, SB Cho - … of the 9th International Conference on …, 2002 - ieeexplore.ieee.org
… Using a special characteristic of SVM-classifier, this paper proposes the algorithm that
improves the performance of text classification by incremental addition of unlabeled data to the …

Semi-superyised support vector machines for unlabeled data classification

G Fung, OL Mangasarian - Optimization methods and software, 2001 - Taylor & Francis
classifying unlabeled data based on the following ideas: (i) A small representative percentage
(5% to 10%) of the unlabeled data … remaining bulk of the unlabeled dataset to one of two …

[PDF][PDF] Learning to classify texts using positive and unlabeled data

X Li, B Liu - IJCAI, 2003 - Citeseer
… We use Roc-SVM and Roc-Clu-SVM to denote the classification techniques that employ
Rocchio and Rocchio with clustering to extract reliable negative set respectively (both methods …

Labeled and unlabeled data in text categorization

C Silva, B Ribeiro - … Joint Conference on Neural Networks (IEEE …, 2004 - ieeexplore.ieee.org
… use of unlabeled data as a way to improve classification performance in text categorization.
The ready availability of this kind of data in most … Our objective is to compare, using a SVM

Towards making unlabeled data never hurt

YF Li, ZH Zhou - IEEE transactions on pattern analysis and …, 2014 - ieeexplore.ieee.org
… [16] discussed the reason why unlabeled data can increase classification error for generative
meth… of the inductive SVM. Otherwise we will rely on the prediction of the inductive SVM. …