English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

In many application domains such as medicine, information retrieval,cybersecurity, social media, etc., datasets used for inducing classificationmodels often have an unequal distribution of the instances of each class. Thissituation, known as imbalanced data classification, causes low predictiveperformance for the minority class examples. Thus, the prediction model isunreliable although the overall model accuracy can be acceptable. Oversamplingand undersampling techniques are well-known strategies to deal with thisproblem by balancing the number of examples of each class. However, theireffectiveness depends on several factors mainly related to data intrinsiccharacteristics, such as imbalance ratio, dataset size and dimensionality,overlapping between classes or borderline examples. In this work, the impact ofthese factors is analyzed through a comprehensive comparative study involving40 datasets from different application areas. The objective is to obtain modelsfor automatic selection of the best resampling strategy for any dataset basedon its characteristics. These models allow us to check several factorssimultaneously considering a wide range of values since they are induced fromvery varied datasets that cover a broad spectrum of conditions. This differsfrom most studies that focus on the individual analysis of the characteristicsor cover a small range of values. In addition, the study encompasses both basicand advanced resampling strategies that are evaluated by means of eightdifferent performance metrics, including new measures specifically designed forimbalanced data classification. The general nature of the proposal allows thechoice of the most appropriate method regardless of the domain, avoiding thesearch for special purpose techniques that could be valid for the target data.

Selecting the suitable resampling strategy for imbalanced data classification regarding dataset properties