A Machine Learning Approach to Predict Air Quality in California

Figure 9

(a) Carbon monoxide forecast detail between the period of the 1st of January and the 15th. Both forecasting methods missed to forecast some of the spikes registered in the series in the beginning of January. (b, c) Two scatter plots with model’s respective errors plotted against observed CO values for the same period.

(a)

(b)

(c)