Exposure assessment models for elemental components of particulate matter in an urban environment: A comparison of regression and random forest approaches

Abstract

Exposure assessment for elemental components of particulate matter (PM) using land use modeling is a complex problem due to the high spatial and temporal variations in pollutant concentrations at the local scale. Land use regression (LUR) models may fail to capture complex interactions and non-linear relationships between pollutant concentrations and land use variables. The increasing availability of big spatial data and machine learning methods present an opportunity for improvement in PM exposure assessment models. In this manuscript, our objective was to develop a novel land use random forest (LURF) model and compare its accuracy and precision to a LUR model for elemental components of PM in the urban city of Cincinnati, Ohio. PM smaller than 2.5 μm (PM2.5) and eleven elemental components were measured at 24 sampling stations from the Cincinnati Childhood Allergy and Air Pollution Study (CCAAPS). Over 50 different predictors associated with transportation, physical features, community socioeconomic characteristics, greenspace, land cover, and emission point sources were used to construct LUR and LURF models. Cross validation was used to quantify and compare model performance. LURF and LUR models were created for aluminum (Al), copper (Cu), iron (Fe), potassium (K), manganese (Mn), nickel (Ni), lead (Pb), sulfur (S), silicon (Si), vanadium (V), zinc (Zn), and total PM2.5 in the CCAAPS study area. LURF utilized a more diverse and greater number of predictors than LUR and LURF models for Al, K, Mn, Pb, Si, Zn, TRAP, and PM2.5 all showed a decrease in fractional predictive error of at least 5% compared to their LUR models. LURF models for Al, Cu, Fe, K, Mn, Pb, Si, Zn, TRAP, and PM2.5 all had a cross validated fractional predictive error less than 30%. Furthermore, LUR models showed a differential exposure assessment bias and had a higher prediction error variance. Random forest and other machine learning methods may provide more accurate exposure assessment.

Modeling PM2.5 Urban Pollution Using Machine Learning and Selected Meteorological Parameters

“…Furthermore, we intend to improve our CGM and use it to classify outliers and find their cause. Considering the diverse machine learning models used in air quality prediction, such as Neural Network [13][14][15], regression [18], decision trees, and Support Vector Machine [17], we applied and tested most of these classifiers in this study. Alternative approaches to improve the accuracy of our model would consist of performing a prediction based on an ensemble of different algorithms of data processing and modeling [16,17,22].…”

“…This is the reason why it is increasingly used to predict air quality [13,[17][18][19][20][21]. However, the data mining does not only differ from one study to another, in terms of classification algorithms, but also regarding the used features.…”

“…However, the data mining does not only differ from one study to another, in terms of classification algorithms, but also regarding the used features. Some of them consider a quite exhaustive list of meteorological factors [15,16], whereas others proceed with a careful selection [13,14,17,22] or do not even use climatic parameters at all [18]. Since machine learning is a very promising method to forecast pollution, we propose applying this approach to predict PM 2.5 concentration in Quito.…”

Mining Public Datasets for Modeling Intra-City PM2.5 Concentrations at a Fine Spatial Resolution

“…These air monitoring stations also exist in many other countries. Scientists and government agencies use measurement data from these stations to build and validate air quality models (AQMs) to explain and predict the past and future air pollution levels for unmonitored locations (e.g., [1, 3, 13, 15, 16, 19, 20, 24, 25, 28, 29]). Predictions from these models can then be used to study the associations between long-term air pollution exposure and health impact at finer spatial scales (than simply using the monitored data) [26, 27].…”

Ensemble based J48 and random forest based C6H6 air pollution detection

“…The proposed sensor combination subsystem has been chosen for exploiting both single sensor specificity and situation related connections. Vlachokostas et al (2011) [9] have discussed that there exist steady relationship between traffic-related air contamination and respiratory symptoms. Be that as it may, numerous urban regions are depicted by the nonappearance of the vital observing foundation, particularly for benzene (C6H6), which is a known human cancer-causing agent.…”