Data Preprocessing and Data Preparation

Data preparation and data exploration are important steps in the preparation of data for machine learning applications. These measures ensure that data is of sufficiently high quality for the subsequent training of ML models. The following work content will be carried out across the SPP:

• Definition and validation of methods for assessing existing data quality
• Definition and validation of data pre-processing and preparation processes for the use case of forming processes
• Automation of data processing steps through to the structuring and storage of sensor data in databases or file formats in an efficient manner
• Development of visualisation techniques and feature engineering methods for multidimensional problems in forming technology
• Compilation of best practices in a methodical "toolbox" & AutoML techniques
• Pattern recognition / clustering / knowledge discovery
• Comparison of different methods: NN/PCA/...
• Supervised vs. unsupervised
• Interface to AK4