What is underfitting in machine learning and how can you prevent it?

Underfitting occurs when a machine learning model is too simple or insufficiently trained to capture the patterns in the data. It results in poor performance on both the training and test datasets. Common causes include low model complexity, inadequate training, excessive regularization, and poor feature selection. It can be prevented by using a more expressive model, training for longer, improving feature engineering, reducing unnecessary regularization, and ensuring high-quality training data. The goal is to build a model that learns the true patterns in the data without becoming so complex that it overfits.