Emerging Trends in Classification with Imbalanced Datasets: A Bibliometric Analysis of Progression

Abstract

Imbalanced or unbalanced datasets are defined as the highly skewed distribution of target variable in the field of machine learning. Imbalanced datasets have greatly caught the attention of researchers due to their negative effect on machine learning models in the last decade. Researchers develop various solutions to the problems of imbalanced datasets and contribute to the literature.The increasing number of articles makes it difficult to follow the literature. Review articles contribute to the solution of this problem. The goal of this study is to conduct a bibliometric analysis to find solutions for classification with imbalanced datasets. Bibliometric analysis is a quantitative technique based on extracting statistics from databases. This work is the first bibliometric analysis to address the problem of imbalanced datasets. In this study, data on imbalanced datasets were obtained from the Scopus database with the R Bibliometrix package version 3.1.4, and recent studies and new approaches were summarized. Data on 16255 publications between 1957-2021 were collected by using selected keywords. This collection mainly comprises 8871 articles, 6987 conference papers, and 175 reviews with 1, 66 average citations per year per document. Among the most cited countries, the United States has 106139 total citations followed by China with 13839 citations and Germany has 9524 citations.