The odor emissions generated by treatment plants imply complex envi- ronmental and economic issues. The modern instrumental odor monitoring systems based on array of several sensors, continuously record the gaseous compounds, but they are characterized by poor selectivity thus compromise the possibility to discriminate and identify the emission sources. In this paper, the ability of odor sensors to distinguish the treatment plant sections generating the gaseous compounds is evaluated by a machine learning classification approach, the Random Forest. The goodness of this method is highlighted through apt performance measures and also with respect to the classical multiple discriminant analysis.
BOOK OF SHORT PAPERS: IES 2023
Distefano V.;Palma M.;De Iaco S.;Mazuruse G.
2023-01-01
Abstract
The odor emissions generated by treatment plants imply complex envi- ronmental and economic issues. The modern instrumental odor monitoring systems based on array of several sensors, continuously record the gaseous compounds, but they are characterized by poor selectivity thus compromise the possibility to discriminate and identify the emission sources. In this paper, the ability of odor sensors to distinguish the treatment plant sections generating the gaseous compounds is evaluated by a machine learning classification approach, the Random Forest. The goodness of this method is highlighted through apt performance measures and also with respect to the classical multiple discriminant analysis.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
