A0191
Title: Impact of rarity level and resampling techniques on machine learning classification performance
Authors: Olcay Alpay - Sinop University (Turkey) [presenting]
Abstract: The classification of binary events is frequently discussed in the literature. However, in cases where the distribution of the event of interest is unbalanced, such as with rare events, machine learning algorithms may produce biased results. The classification performance of several machine learning algorithms is investigated, including Logistic Regression, Support Vector Machine, Random Forest, Gradient Boosting, and Artificial Neural Network, in different rarity scenarios and how they are affected by unbalanced data. The impact of resampling techniques on dataset balancing and classification is also examined.
