A High-Frequency Data-Driven Machine Learning Approach for Demand Forecasting in Smart Cities

Figure 14

Example of a good forecast with missing data at the high-density population area site. There were missing/wrong data from approximately 1:00 a.m. until 6:00 a.m. Nevertheless, the overall outcome of the forecast was very accurate. The pointwise 95% confidence band of the estimation is depicted in grey.