Random Balance: Ensembles of variable priors classifiers for imbalanced data

“…The experimental set-up has been designed following the common practice (Akbani et al, 2004;Barua et al, 2014;Díez-Pastor et al, 2015;Galar et al, 2013;López et al, 2012López et al, , 2013Yu et al, 2013Yu et al, , 2015, which consists of generating skewed data sets with different levels of class imbalance. In particular, the scheme proposed in the literature is to transform multi-class data sets by combining several original classes to shape the majority and minority classes.…”

Section: Data Setsmentioning

confidence: 99%

“…While research in class imbalance has mainly concentrated on well-known classifiers such as support vector machines (Akbani et al, 2004;Hwang et al, 2011;Liu et al, 2011;Maldonado and López, 2014;Yu et al, 2015), kernel methods (Hong et al, 2007;Maratea et al, 2014), k-nearest neighbors (Dubey and Pudi, 2013), decision trees (Kang and Ramamohanarao, 2014) and multiple classifier systems (Díez-Pastor et al, 2015;Galar et al, 2012;Krawczyk et al, 2014;Park and Ghosh, 2014), very few theoretical or empirical analyses have been done so far to thoroughly establish the performance of associative memories when learning from class imbalanced data. Therefore, the present study intends to extend the very preliminary existing works (Cleofas-Sánchez et al, 2013 by increasing the scope and detail at which a type of associative memory networks (the hybrid associative classifier with translation) performs in the framework of class imbalance.…”

Section: Introductionmentioning

confidence: 99%

Associative learning on imbalanced environments: An empirical study

Cleofas-Sánchez

Sánchez

García

et al. 2016

Expert Systems with Applications

View full text Add to dashboard Cite

Associative memories have emerged as a powerful computational neural network model for several pattern classification problems. Like most traditional classifiers, these models assume that the classes share similar prior probabilities. However, in many real-life applications the ratios of prior probabilities between classes are extremely skewed. Although the literature has provided numerous studies that examine the performance degradation of renowned classifiers on different imbalanced scenarios, so far this effect has not been supported by a thorough empirical study in the context of associative memories. In this paper, we fix our attention on the applicability of the associative neural networks to the classification of imbalanced data. The key questions here addressed are whether these models perform better, the same or worse than other popular classifiers, how the level of imbalance affects their performance, and whether distinct resampling strategies produce a different impact on the associative memories. In order to answer these questions and gain further insight into the feasibility and efficiency of the associative memories, a large-scale experimental evaluation with 31 databases, seven classification models and four resampling algorithms is carried out here, along with a non-parametric statistical test to discover any significant differences between each pair of classifiers.

show abstract

“…Although it is related to undersampling methods, it takes a step away from them by allowing the removal of minority elements from the dataset, as in [25]. Most existing methods do not allow such kind of reduction of non-representative or noisy elements from the positive class.…”

Section: Introductionmentioning

confidence: 99%

EPRENNID: An evolutionary prototype reduction based ensemble for nearest neighbor classification of imbalanced data

et al. 2016

View full text Add to dashboard Cite

Classification problems with an imbalanced class distribution have received an increased amount of attention within the machine learning community over the last decade. They are encountered in a growing number of realworld situations and pose a challenge to standard machine learning techniques. We propose a new hybrid method specifically tailored to handle class imbalance, called EPRENNID. It performs an evolutionary prototype reduction focused on providing diverse solutions to prevent the method from overfitting the training set. It also allows us to explicitly reduce the underrepresented class, which the most common preprocessing solutions handling class imbalance usually protect. As part of the experimental study, we show that the proposed prototype reduction method outperforms state-of-the-art preprocessing techniques. The preprocessing step yields multiple prototype sets that are later used in an ensemble, performing a weighted voting scheme with the nearest neighbor classifier. EPRENNID is experimentally shown to significantly outperform previous proposals.

show abstract

Random Balance: Ensembles of variable priors classifiers for imbalanced data

Cited by 205 publications

References 43 publications

The study of under- and over-sampling methods’ utility in analysis of highly imbalanced data on osteoporosis

The study of under- and over-sampling methods’ utility in analysis of highly imbalanced data on osteoporosis

Associative learning on imbalanced environments: An empirical study

EPRENNID: An evolutionary prototype reduction based ensemble for nearest neighbor classification of imbalanced data

Contact Info

Product

Resources

About