Use of Ensemble Based on GA for Imbalance Problem

摘要: In real-world applications, it has been observed that class imbalance (significant differences in prior probabilities) may produce an important deterioration of the classifier performance, particular with patterns belonging to less represented classes. One method tackle this problem consists resample original training set, either by over-sampling minority and/or under-sampling majority class. paper, we propose two ensemble models (using a modular neural network and nearest neighbor rule) trained on datasets under-sampled genetic algorithms. Experiments real demonstrate effectiveness methodology here proposed.