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

<div>Forecasting at the high-density population area on Thursday, 24 July 2014. The water flow jumped at 10:28 a.m. from 9.39 l/min to 19.03 l/min. This offset of around 10 l/min was maintained after the jump. Therefore, although the RMSE was not very high (RMSE = 6.81 l/min), the FOB put that day in red (FOB = 0.63). The pointwise 95% confidence band of the estimation is depicted in grey.</div>

Scientific Programming

fig9

Figure 9

Figure 9: A High-Frequency Data-Driven Machine Learning Approach for Demand Forecasting in Smart Cities 