Under-Sampling Approaches for Improving Prediction of the Minority Class in an Imbalanced Dataset

Yen, Show‐Jane; Lee, Yue‐Shi

doi:10.1007/978-3-540-37256-1_89

Cited by 104 publications

(25 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Alternative approaches that could help balancing datasets include undersampling of the bigger class [32,33] or oversampling of the smaller class. [34] The extended discussion of these approaches is beyond the scope of this review.…”

Section: Dataset Size and Balancingmentioning

confidence: 99%

Best Practices for QSAR Model Development, Validation, and Exploitation

Tropsha

2010

Molecular Informatics

1,486

1,150

View full text Add to dashboard Cite

After nearly five decades "in the making", QSAR modeling has established itself as one of the major computational molecular modeling methodologies. As any mature research discipline, QSAR modeling can be characterized by a collection of well defined protocols and procedures that enable the expert application of the method for exploring and exploiting ever growing collections of biologically active chemical compounds. This review examines most critical QSAR modeling routines that we regard as best practices in the field. We discuss these procedures in the context of integrative predictive QSAR modeling workflow that is focused on achieving models of the highest statistical rigor and external predictive power. Specific elements of the workflow consist of data preparation including chemical structure (and when possible, associated biological data) curation, outlier detection, dataset balancing, and model validation. We especially emphasize procedures used to validate models, both internally and externally, as well as the need to define model applicability domains that should be used when models are employed for the prediction of external compounds or compound libraries. Finally, we present several examples of successful applications of QSAR models for virtual screening to identify experimentally confirmed hits.

show abstract

Section: Dataset Size and Balancingmentioning

confidence: 99%

Best Practices for QSAR Model Development, Validation, and Exploitation

Tropsha

2010

Molecular Informatics

1,486

1,150

View full text Add to dashboard Cite

show abstract

“…Besides, if a positive instance is misclassified by its three nearest neighbors, then the algorithm also eliminates the neighbors that belong to the majority class. A quite different alternative corresponds to under-Sampling Based on Clustering (SBC) [22], which rests on the idea that there may exist different clusters in a given data set, and each cluster may have distinct characteristics depending on the ratio of the number of positive examples to the number of negative examples in the cluster. Thus the SBC algorithm first gathers all examples in the data set into some clusters, and then determines the number of negative cases that will be randomly picked up.…”

Section: Under-samplingmentioning

confidence: 99%

Improving Risk Predictions by Preprocessing Imbalanced Credit Data

García

Marqués

Sánchez

2012

Neural Information Processing

View full text Add to dashboard Cite

Abstract. Imbalanced credit data sets refer to databases in which the class of defaulters is heavily under-represented in comparison to the class of non-defaulters. This is a very common situation in real-life credit scoring applications, but it has still received little attention. This paper investigates whether data resampling can be used to improve the performance of learners built from imbalanced credit data sets, and whether the effectiveness of resampling is related to the type of classifier. Experimental results demonstrate that learning with the resampled sets consistently outperforms the use of the original imbalanced credit data, independently of the classifier used.

show abstract

“…This occurs when there are many more instances in one class than others in a training collection. In this case, a classifier usually tends to predict samples from the majority class and completely ignore the minority class [37]. For this reason, we applied an approach to subsampling (reducing the number of samples that have the majority class) to generate a balanced collection composed of:…”

Section: Features Selection Algorithms Studymentioning

confidence: 99%