In a classification problem, you have trained multiple models using ensemble techniques. You want to combine their predictions to improve the overall performance. Which of the following methods can be used for combining the predictions, and explain how each method works?