AutoML stands for "Automated Machine Learning".
When using Machine Learning, you can ask yourself, "Which model is better, such as regression analysis or neural network?" There is a trial and error process of "What is the best feature engineering?" The general procedure is fixed. The procedure is to "try the model from the simple" and "try the feature from the one that does not process anything".
AutoML is software that automates that process.
AutoML is a way to create models with high prediction accuracy, so the first use is prediction.
The second use is to compress the high dimension into two dimensions in a regression analysis system and use it as a visualization method to heuristically perform Data Mining, It is also possible to use it as a hint for the Quantity Way of Making Hypothesis.
R has packages called automl and h2o. AutoML is a tool for automatic parameter adjustment of Neural Network. Just specify the Y and X variables to make it work. However, for data that can be created neatly by ordinary regression analysis, I make a model with a large error, and I have confirmed that it takes a lot of time. It's a relatively new package published in 2020, so we're waiting and seeing. In the Visualization by compressing high-dimensional into two dimensions with regression analysis system by R, automl can be used.
H2O is difficult to configure, and even simple data gives errors, and I have not been able to produce results.
There's PyCaret, and the most common models have the ability to try them out and show you the results.
NEXT 
Bayesian Network
