In order to use today’s powerful machine learning systems even without detailed machine learning knowledge, e.g., the meaning and effects of the involved hyperparameters, tools for automating the entire process of data analysis and machine learning are necessary. At the very least, one would like to assist a human analyst with the help of machine tools to process the individual steps. The aim of this core project is the further development of methods from the area of AutoML. AutoML is looking for automatic or semi-automatic methods for (1.) data preparation and transformation, (2.) selection of suitable analysis algorithms, (3.) execution of the algorithm, (4.) determination of good hyperparameters of the machine learning algorithms, (5.) selection of measures of success, (6.) data collection of missing data, and (7.) report generation.
