Which of the following is done by at least one of the ensemble modeling methods? Select all that apply.
1. The method makes a series of models, with each model intended to improve on the last one by adding to the data set extra copies of the data points that the last model predicted incorrectly.
2. The method improves on the model by automatically collecting additional data that are more recent than the data included in the original data set.
3. The method makes its predictions by combining the predictions of multiple models, each made with a different algorithm.
4. The method divides the data into three subsets instead of two so that models can be tested against a new subset of data in both the tuning round and the test round.