Page 72 - Contributed Paper Session (CPS) - Volume 7
P. 72
CPS2031 Javier Linkolk L. et al.
Figure 1: SOM ensemble of LSTM
3. Result
We present the results of a neural network SOM+LSTM model designed for
the characterization and estimation of hourly pollution PM2.5 concentrations
by one station in the Metropolitan area: "La Florida".
Figure 2: Time series of PM2.5 “La Florida” Figure 3: SOM of the data
In figure 2, the behaviour of the contamination data of the "La Florida" station
during the 365 days is reported. The SOM is used to group the time series
according to the pattern. Figure 3 reports neurons that have a group of similar
time series corresponding to the pollution record per hour in a day.
59 | I S I W S C 2 0 1 9
