Changing data distributions are a major challenge in time series forecasting. To address this problem, we present three novel online approaches that lead to self-adaptive Gaussian Processes. We further assess time series forecasting methods for predicting sales of small and medium-sized companies dealing with perishable goods despite distribution shifts. In addition, we provide a programming framework to ensure reproducible and comparable results as well as easy access to new methods.      «

Changing data distributions are a major challenge in time series forecasting. To address this problem, we present three novel online approaches that lead to self-adaptive Gaussian Processes. We further assess time series forecasting methods for predicting sales of small and medium-sized companies dealing with perishable goods despite distribution shifts. In addition, we provide a programming framework to ensure reproducible and comparable results as well as easy access to new methods....     »