Test architecture of the proposed system. Paulo S. G. de Mattos Neto George D. C. Cavalcanti Francisco Madeiro Tiago A. E. Ferreira 10.1371/journal.pone.0138507.g003 https://plos.figshare.com/articles/figure/_Test_architecture_of_the_proposed_system_/1558899 <p>Test architecture of the proposed system.</p> 2015-09-28 03:21:04 ga Artificial neural networks approach correction PM 10 concentration series High PM concentration hs residuals modeling fitness function ann PM forecasters forecasting method performance Experimental results show PM time series