BFS - Plausibility check on quality and reliability of administrative data

This project intends to extend and accelerate the plausibility checks in the Federal Office for Statistics (BFS) using machine learning algorithms while increasing the data quality. Data that appears to be either clearly wrong or at least questionable will be sent back to the data providers with a request for correction or comment. Until now, such plausibility checks were mostly carried out in two different ways: either by manual controls or by automated procedures using limit values ​​and logical tests. This process of mutual plausi checks is very labor intensive. Machine learning can help to carry out the controls more quickly and more precisely.