Epoch

An epoch in machine learning is one complete cycle through the entire set of training data. During an epoch, the model makes predictions for each data point, calculates errors, and updates its internal parameters to improve accuracy. Multiple epochs allow the model to learn patterns more effectively, refining its predictions with each pass. Think of it like reviewing your notes multiple times to better understand the material; each pass helps the model get closer to accurate predictions. The goal is to find the right number of epochs that lead to good learning without overfitting.