A Janela de Tempo Ideal: Otimizando Indicadores Bibliométricos para Prever o Sucesso de Curto Prazo de Pesquisadores

Abstract

Predicting the success of a researcher by evaluating the success of their publications is an important topic that is attracting more and more attention in the scientific community. However, in this context, there are no comprehensive studies that provide the ideal window size for capturing publications as a basis for calculating predictors. In this study, we are interested in how short or long time windows used as the basis for calculating predictive bibliometric indices affect the accuracy of the classifier that categorizes researchers into those with successful recent publications and other researchers. Using the American Physical Society (APS) dataset, we compare the performance of the classifiers. We use 10- fold cross-validation to determine the most accurate classifier in both scenarios described above. We found that a more comprehensive evaluation of scientists is necessary, suggesting that a long window is superior to a short window.