Results 1  10
of
3,415
Wrappers for Feature Subset Selection
 AIJ SPECIAL ISSUE ON RELEVANCE
, 1997
"... In the feature subset selection problem, a learning algorithm is faced with the problem of selecting a relevant subset of features upon which to focus its attention, while ignoring the rest. To achieve the best possible performance with a particular learning algorithm on a particular training set, a ..."
Abstract

Cited by 1522 (3 self)
 Add to MetaCart
In the feature subset selection problem, a learning algorithm is faced with the problem of selecting a relevant subset of features upon which to focus its attention, while ignoring the rest. To achieve the best possible performance with a particular learning algorithm on a particular training set, a feature subset selection method should consider how the algorithm and the training set interact. We explore the relation between optimal feature subset selection and relevance. Our wrapper method searches for an optimal feature subset tailored to a particular algorithm and a domain. We study the strengths and weaknesses of the wrapper approach andshow a series of improved designs. We compare the wrapper approach to induction without feature subset selection and to Relief, a filter approach to feature subset selection. Significant improvement in accuracy is achieved for some datasets for the two families of induction algorithms used: decision trees and NaiveBayes.
An introduction to variational methods for graphical models
 TO APPEAR: M. I. JORDAN, (ED.), LEARNING IN GRAPHICAL MODELS
"... ..."
Probabilistic Outputs for Support Vector Machines and Comparisons to Regularized Likelihood Methods
 ADVANCES IN LARGE MARGIN CLASSIFIERS
, 1999
"... The output of a classifier should be a calibrated posterior probability to enable postprocessing. Standard SVMs do not provide such probabilities. One method to create probabilities is to directly train a kernel classifier with a logit link function and a regularized maximum likelihood score. Howev ..."
Abstract

Cited by 1041 (0 self)
 Add to MetaCart
(Show Context)
The output of a classifier should be a calibrated posterior probability to enable postprocessing. Standard SVMs do not provide such probabilities. One method to create probabilities is to directly train a kernel classifier with a logit link function and a regularized maximum likelihood score. However, training with a maximum likelihood score will produce nonsparse kernel machines. Instead, we train an SVM, then train the parameters of an additional sigmoid function to map the SVM outputs into probabilities. This chapter compares classification error rate and likelihood scores for an SVM plus sigmoid versus a kernel method trained with a regularized likelihood error function. These methods are tested on three dataminingstyle data sets. The SVM+sigmoid yields probabilities of comparable quality to the regularized maximum likelihood kernel method, while still retaining the sparseness of the SVM.
A Comparison of Methods for Multiclass Support Vector Machines
 IEEE TRANS. NEURAL NETWORKS
, 2002
"... Support vector machines (SVMs) were originally designed for binary classification. How to effectively extend it for multiclass classification is still an ongoing research issue. Several methods have been proposed where typically we construct a multiclass classifier by combining several binary class ..."
Abstract

Cited by 935 (22 self)
 Add to MetaCart
(Show Context)
Support vector machines (SVMs) were originally designed for binary classification. How to effectively extend it for multiclass classification is still an ongoing research issue. Several methods have been proposed where typically we construct a multiclass classifier by combining several binary classifiers. Some authors also proposed methods that consider all classes at once. As it is computationally more expensive to solve multiclass problems, comparisons of these methods using largescale problems have not been seriously conducted. Especially for methods solving multiclass SVM in one step, a much larger optimization problem is required so up to now experiments are limited to small data sets. In this paper we give decomposition implementations for two such “alltogether” methods. We then compare their performance with three methods based on binary classifications: “oneagainstall,” “oneagainstone,” and directed acyclic graph SVM (DAGSVM). Our experiments indicate that the “oneagainstone” and DAG methods are more suitable for practical use than the other methods. Results also show that for large problems methods by considering all data at once in general need fewer support vectors.
Improved Boosting Algorithms Using Confidencerated Predictions
 MACHINE LEARNING
, 1999
"... We describe several improvements to Freund and Schapire’s AdaBoost boosting algorithm, particularly in a setting in which hypotheses may assign confidences to each of their predictions. We give a simplified analysis of AdaBoost in this setting, and we show how this analysis can be used to find impr ..."
Abstract

Cited by 933 (26 self)
 Add to MetaCart
We describe several improvements to Freund and Schapire’s AdaBoost boosting algorithm, particularly in a setting in which hypotheses may assign confidences to each of their predictions. We give a simplified analysis of AdaBoost in this setting, and we show how this analysis can be used to find improved parameter settings as well as a refined criterion for training weak hypotheses. We give a specific method for assigning confidences to the predictions of decision trees, a method closely related to one used by Quinlan. This method also suggests a technique for growing decision trees which turns out to be identical to one proposed by Kearns and Mansour. We focus next on how to apply the new boosting algorithms to multiclass classification problems, particularly to the multilabel case in which each example may belong to more than one class. We give two boosting methods for this problem, plus a third method based on output coding. One of these leads to a new method for handling the singlelabel case which is simpler but as effective as techniques suggested by Freund and Schapire. Finally, we give some experimental results comparing a few of the algorithms discussed in this paper.
Boosting the margin: A new explanation for the effectiveness of voting methods
 IN PROCEEDINGS INTERNATIONAL CONFERENCE ON MACHINE LEARNING
, 1997
"... One of the surprising recurring phenomena observed in experiments with boosting is that the test error of the generated classifier usually does not increase as its size becomes very large, and often is observed to decrease even after the training error reaches zero. In this paper, we show that this ..."
Abstract

Cited by 896 (52 self)
 Add to MetaCart
One of the surprising recurring phenomena observed in experiments with boosting is that the test error of the generated classifier usually does not increase as its size becomes very large, and often is observed to decrease even after the training error reaches zero. In this paper, we show that this phenomenon is related to the distribution of margins of the training examples with respect to the generated voting classification rule, where the margin of an example is simply the difference between the number of correct votes and the maximum number of votes received by any incorrect label. We show that techniques used in the analysis of Vapnik’s support vector classifiers and of neural networks with small weights can be applied to voting methods to relate the margin distribution to the test error. We also show theoretically and experimentally that boosting is especially effective at increasing the margins of the training examples. Finally, we compare our explanation to those based on the biasvariance decomposition.
On the optimality of the simple Bayesian classifier under zeroone loss
 MACHINE LEARNING
, 1997
"... The simple Bayesian classifier is known to be optimal when attributes are independent given the class, but the question of whether other sufficient conditions for its optimality exist has so far not been explored. Empirical results showing that it performs surprisingly well in many domains containin ..."
Abstract

Cited by 805 (26 self)
 Add to MetaCart
The simple Bayesian classifier is known to be optimal when attributes are independent given the class, but the question of whether other sufficient conditions for its optimality exist has so far not been explored. Empirical results showing that it performs surprisingly well in many domains containing clear attribute dependences suggest that the answer to this question may be positive. This article shows that, although the Bayesian classifier’s probability estimates are only optimal under quadratic loss if the independence assumption holds, the classifier itself can be optimal under zeroone loss (misclassification rate) even when this assumption is violated by a wide margin. The region of quadraticloss optimality of the Bayesian classifier is in fact a secondorder infinitesimal fraction of the region of zeroone optimality. This implies that the Bayesian classifier has a much greater range of applicability than previously thought. For example, in this article it is shown to be optimal for learning conjunctions and disjunctions, even though they violate the independence assumption. Further, studies in artificial domains show that it will often outperform more powerful classifiers for common training set sizes and numbers of attributes, even if its bias is a priori much less appropriate to the domain. This article’s results also imply that detecting attribute dependence is not necessarily the best way to extend the Bayesian classifier, and this is also verified empirically.
Approximate Statistical Tests for Comparing Supervised Classification Learning Algorithms
, 1998
"... This article reviews five approximate statistical tests for determining whether one learning algorithm outperforms another on a particular learning task. These tests are compared experimentally to determine their probability of incorrectly detecting a difference when no difference exists (type I err ..."
Abstract

Cited by 718 (9 self)
 Add to MetaCart
This article reviews five approximate statistical tests for determining whether one learning algorithm outperforms another on a particular learning task. These tests are compared experimentally to determine their probability of incorrectly detecting a difference when no difference exists (type I error). Two widely used statistical tests are shown to have high probability of type I error in certain situations and should never be used: a test for the difference of two proportions and a paireddifferences t test based on taking several random traintest splits. A third test, a paireddifferences t test based on 10fold crossvalidation, exhibits somewhat elevated probability of type I error. A fourth test, McNemar’s test, is shown to have low type I error. The fifth test is a new test, 5 × 2 cv, based on five iterations of twofold crossvalidation. Experiments show that this test also has acceptable type I error. The article also measures the power (ability to detect algorithm differences when they do exist) of these tests. The crossvalidated t test is the most powerful. The 5×2 cv test is shown to be slightly more powerful than McNemar’s test. The choice of the best test is determined by the computational cost of running the learning algorithm. For algorithms that can be executed only once, McNemar’s test is the only test with acceptable type I error. For algorithms that can be executed 10 times, the 5×2 cv test is recommended, because it is slightly more powerful and because it directly measures variation due to the choice of training set.
Statistical Comparisons of Classifiers over Multiple Data Sets
, 2006
"... While methods for comparing two learning algorithms on a single data set have been scrutinized for quite some time already, the issue of statistical tests for comparisons of more algorithms on multiple data sets, which is even more essential to typical machine learning studies, has been all but igno ..."
Abstract

Cited by 716 (0 self)
 Add to MetaCart
While methods for comparing two learning algorithms on a single data set have been scrutinized for quite some time already, the issue of statistical tests for comparisons of more algorithms on multiple data sets, which is even more essential to typical machine learning studies, has been all but ignored. This article reviews the current practice and then theoretically and empirically examines several suitable tests. Based on that, we recommend a set of simple, yet safe and robust nonparametric tests for statistical comparisons of classifiers: the Wilcoxon signed ranks test for comparison of two classifiers and the Friedman test with the corresponding posthoc tests for comparison of more classifiers over multiple data sets. Results of the latter can also be neatly presented with the newly introduced CD (critical difference) diagrams.