Abstract
The advances within the Internet of Things and sensor systems put the focus on the improvement of the data reliability as close to the edge as possible. This work investigates how well-established techniques can be used for the imputation of contaminated data. We look at the performance of four algorithms for different contamination rates and error bursts of variable length. Furthermore, the algorithms are also evaluated on a constrained environment to showcase the behavior of data imputation methods at the edge of an IoT-based system.