For each type of model you can configure the search space. These include general settings as well as individual hyperparameter ranges for algorithms.
Kfolds, is a cross validation technique for splitting data into train/test. By default 5 is selected.
Scoring Function, the scoring function is used to assess how a model is performing and create a leaderboard of results.
Train Split, this can be used in place of Kfolds and is the percentage of training data that will be used. If left unchanged Kfolds will be used to split data.
Max Trials, this is the maximum number of trials to be run before training stops.
Max Trial Time Minutes, this is the maximum time in minutes an individual trial can run before it is stopped.
Max Total Time Minutes, the maximum time in minutes an entire training can run for before it is stopped.
The list of supported algorithms to be included in the search space. Hyperparameters can also be configured for each algorithm. See Machine Learning Algorithms for more information on each algorithm.