Each Machine Learning algorithm needs to be trained using the data to make proper predictions. This brick is designed to be flexible to all implemented by datrics.ai algorithms.
Firstly, you need to connect the desired ML model and train dataset to the brick. After that, you might need to specify the date and target columns. (Some algorithms like Prophet require both, while unsupervised algorithms require none. Regression and Classification algorithms usually can utilize date column but require only the target variable).
When the training is finished, you may check its performance on the train data by clicking "Model performance":