Untitled

Li, Yang; Long, Philip M.

doi:10.1023/a:1012435301888

Cited by 133 publications

(1 citation statement)

References 36 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…MDLText foi comparado com os seguintes métodos tradicionais de aprendizado online: M.NB[McCallum e Nigam 1998], B.NB[McCallum e Nigam 1998], Perceptron[Freund e Schapire 1999], gradiente descendente estocástico (SGD -Stochastic Gradient Descent)[Zhang 2004], algoritmo de margem larga aproximada (ALMA -Approximate Large Margin Algorithm)[Hoi et al 2014, Gentile 2002, gradiente descendente online (OGD -Online Gradient Descent)[Zinkevich 2003] e algoritmo de margem máxima online relaxada (ROMMA -Relaxed Online Maximum Margin Algorithm)[Li e Long 2002].Os métodos M.NB, B.NB foram utilizados a partir da biblioteca scikit-learn em Python. Os experimentos com os métodos Perceptron, ALMA, OGD, SGD e ROMMA foram realizados usando funções da biblioteca LIBOL 4[Hoi et al 2014] para MATLAB.…”

unclassified

MDLText aplicado na Filtragem Automática de SPIM e SMS Spam

Silva

Almeida²,

Yamakami

2018

iSys

View full text Add to dashboard Cite

A filtragem automática de spam em mensagens instantâneas e SMS é um problema desafiador, pois as mensagens são frequentemente curtas e repletas de ruídos, tais como gírias, expressões idiomáticas, símbolos, emoticons e abreviações, o que dificulta a extração de conhecimento e predição. Para enfrentar esse problema, neste artigo é avaliado um método de classificação de texto baseado no princípio da descrição mais simples, que é eficiente, rápido, escalável, multiclasse e possui aprendizado incremental. Experimentos realizados com uma base de dados real e pública, em cenários de aprendizado online e offline, indicam que o método proposto é promissor para a tarefa de detecção de spam em mensagens instantâneas e SMS.

show abstract

unclassified

MDLText aplicado na Filtragem Automática de SPIM e SMS Spam

Silva

Almeida²,

Yamakami

2018

iSys

View full text Add to dashboard Cite

show abstract

A family of large margin linear classifiers and its application in dynamic environments

Shen

Dietterich

2009

Statistical Analysis

View full text Add to dashboard Cite

Real-time problems, in which the learning must be fast and the importance of the features might be changing, pose a challenge to machine learning algorithms. To learn robust classifiers in such nonstationary environments, it is essential not to assign too much weight to any single feature. We solve the problems by combining regularization mechanisms with online large margin learning algorithms. We prove bounds on their error and show that removing features with small weights has little influence on the accuracy, suggesting that these methods exhibit feature selection ability. We show that such regularized learning algorithms automatically decrease the influence of the old training instances and focus on the more recent ones. This makes them especially attractive in the dynamic environments. We present experimental results on real datasets to show the merit of our algorithms.

show abstract

Online training on a budget of support vector machines using twin prototypes

Wang

Vučetić

2010

Statistical Analysis

View full text Add to dashboard Cite

This paper proposes twin prototype support vector machine (TVM), a constant space and sublinear time support vector machine (SVM) algorithm for online learning. TVM achieves its favorable scaling by memorizing only a fixed-size data summary in the form of example prototypes and their associated information during training. In addition, TVM guarantees that the optimal SVM solution is maintained on all prototypes at any time. To maximize the accuracy of TVM, prototypes are constructed to approximate the data distribution near the decision boundary. Given a new training example, TVM is updated in three steps. First, the new example is added as a new prototype if it is near the decision boundary. If this happens, to maintain the budget, either the prototype farthest away from the decision boundary is removed or two near prototypes are selected and merged into a single one. Finally, TVM is updated by incremental and decremental techniques to account for the change. Several methods for prototype merging were proposed and experimentally evaluated. TVM algorithms with hinge loss and ramp loss were implemented and thoroughly tested on 12 large datasets. In most cases, the accuracy of low-budget TVMs was comparable with the resource-unconstrained SVMs. Additionally, the TVM accuracy was substantially larger than that of SVM trained on a random sample of the same size. Even larger difference in accuracy was observed when comparing with Forgetron, a popular budgeted kernel perceptron algorithm. As expected, the difference in accuracy between hinge loss and ramp loss TVM was negligible and hinge loss version is preferable due to its lower computational cost. The results illustrate that highly accurate online SVMs could be trained from arbitrary large data streams using devices with severely limited memory budgets.

show abstract

Untitled

Cited by 133 publications

References 36 publications

MDLText aplicado na Filtragem Automática de SPIM e SMS Spam

MDLText aplicado na Filtragem Automática de SPIM e SMS Spam

A family of large margin linear classifiers and its application in dynamic environments

Online training on a budget of support vector machines using twin prototypes

Contact Info

Product

Resources

About