Bagging (Bootstrap Aggregating) is a popular ensemble learning method that enhances the performance and resilience of machine learning models by combining predictions from multiple models trained on random subsets of a dataset. This approach minimizes variance and improves generalization.

In this project, we utilized CNNs with bagging to classify the CIFAR-100 dataset, a benchmark dataset containing 60,000 color images across 100 classes. Of these, 50,000 images are used for training, and 10,000 for testing. Each image has a resolution of 32×32 pixels, presenting challenges due to high intra-class variability and low resolution.