An Empirical Study for the Multi-class Imbalance Problem with Neural Networks

Alejo, R.; Sotoca, José Martínez; Casañ, Gustavo A.

doi:10.1007/978-3-540-85920-8_59

Cited by 12 publications

(9 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Machine learning community suggests many tricks to solve the deficiency of its models in predicting the small classes like U2R and R2L classes of ID problem, it is the problem mentioned in the previous review. Different approaches suggested [14,15] to solve the imbalanced classes like resampling techniques and algorithmic approach. The oversampling and undersampling are the common two resampling methods in literature while the cost function was added to different machine learning algorithms to address it's sensitivity to imbalanced classes.…”

Section: Related Workmentioning

confidence: 99%

“…The oversampling and undersampling are the common two resampling methods in literature while the cost function was added to different machine learning algorithms to address it's sensitivity to imbalanced classes. Different cost functions suggested in both works and applied with Neural Networks [15] and Extreme machine learning (ELM) [14] algorithms. This improves the prediction for the small classes in all used data set.…”

Section: Related Workmentioning

confidence: 99%

“…SVM, as mentioned, is a powerful and robust classification algorithm, but it is like most pattern recognition models make accuracy paradox when training using imbalanced classes. Cost-function is one of the methods were suggested to address this problem [15] .It gives different weights to the records that belong to different classes, many cost-function options were suggested by Alejo, et al [15], the option 2 was used here which illustrated in the following equation:…”

Section: Accuracy Paradox and Cost-function Schemamentioning

confidence: 99%

See 2 more Smart Citations

Using weighted Support Vector Machine to address the imbalanced classes problem of Intrusion Detection System

Alabdallah¹,

Awad²

2018

KSII TIIS

View full text Add to dashboard Cite

Improving the intrusion detection system (IDS) is a pressing need for cyber security world. With the growth of computer networks, there are constantly daily new attacks. Machine Learning (ML) is one of the most important fields which have great contribution to address the intrusion detection issues. One of these issues relates to the imbalance of the diverse classes of network traffic. Accuracy paradox is a result of training ML algorithm with imbalanced classes. Most of the previous efforts concern improving the overall accuracy of these models which is truly important. However, even they improved the total accuracy of the system; it fell in the accuracy paradox. The seriousness of the threat caused by the minor classes and the pitfalls of the previous efforts to address this issue is the motive for this work. In this paper, we consolidated stratified sampling, cost function and weighted Support Vector Machine (WSVM) method to address the accuracy paradox of ID problem. This model achieved good results of total accuracy and superior results in the small classes like the User-To-Remote and Remote-To-Local attacks using the improved version of the benchmark dataset KDDCup99 which is called NSL-KDD.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Accuracy Paradox and Cost-function Schemamentioning

confidence: 99%

See 1 more Smart Citation

Using weighted Support Vector Machine to address the imbalanced classes problem of Intrusion Detection System

Alabdallah¹,

Awad²

2018

KSII TIIS

View full text Add to dashboard Cite

show abstract

“…Some learning algorithms based on cost-errors, such as those in [29], have been proposed and shown to be efficient in dealing with imbalanced classification problems. So in this paper, in order to correct the learning deviation of the widths, we will also adopt the cost-errors using the count ratio of the samples of each region to modify the objective function in (2).…”

Section: The Imbalanced Problemmentioning

confidence: 99%

Optimizing widths with PSO for center selection of Gaussian radial basis function networks

Zhao

et al. 2013

Sci. China Inf. Sci.

View full text Add to dashboard Cite

The radial basis function (RBF) centers play different roles in determining the classification capability of a Gaussian radial basis function neural network (GRBFNN) and should hold different width values. However, it is very hard and time-consuming to optimize the centers and widths at the same time. In this paper, we introduce a new insight into this problem. We explore the impact of the definition of widths on the selection of the centers, propose an optimization algorithm of the RBF widths in order to select proper centers from the center candidate pool, and improve the classification performance of the GRBFNN. The design of the objective function of the optimization algorithm is based on the local mapping capability of each Gaussian RBF. Further, in the design of the objective function, we also handle the imbalanced problem which may occur even when different local regions have the same number of examples. Finally, the recursive orthogonal least square (ROLS) and genetic algorithm (GA), which are usually adopted to optimize the RBF centers, are separately used to select the centers from the center candidates with the initialized widths, in order to testify the validity of our proposed width initialization strategy on the selection of centers. Our experimental results show that, compared with the heuristic width setting method, the width optimization strategy makes the selected centers more appropriate, and improves the classification performance of the GRBFNN. Moreover, the GRBFNN constructed by our method can attain better classification performance than the RBF LS-SVM, which is a state-of-the-art classifier.

show abstract

“…Classification problems that breach this tacit assumption are ill suited to such algorithms (Alejo et al, 2008). The learning objectives of these algorithms focus on classification accuracy of the training data set; under-sampled classes are poorly represented during mean-squared error like decision rules and the resultant classification rule becomes unduly biased (Sun et al, 2009).…”

Section: Introductionmentioning

confidence: 99%

Feature extraction for novelty detection as applied to fault detection in machinery

McBain

Timusk

2011

Pattern Recognition Letters

View full text Add to dashboard Cite

An Empirical Study for the Multi-class Imbalance Problem with Neural Networks

Cited by 12 publications

References 8 publications

Using weighted Support Vector Machine to address the imbalanced classes problem of Intrusion Detection System

Using weighted Support Vector Machine to address the imbalanced classes problem of Intrusion Detection System

Optimizing widths with PSO for center selection of Gaussian radial basis function networks

Feature extraction for novelty detection as applied to fault detection in machinery

Contact Info

Product

Resources

About