The impact of data set similarity and diversity on transfer learning success in time series forecasting