Table 6 Summary of studies forecasting air pollution distributions and related variables using data mining methods

Kolehmainen et al. [60]

NO2

ANN

Comparing two Neural Nets for their suitability in forecasting Air Quality

Kukkonen et al. [33]

PM NO₂,

ANN

Machine Learning Model comparison for forecasting NO2 and PM10 concentrations

Niska et al. [59]

NO2

Genetic Algorithms, ANN

Investigate the use of GA to find a better ANN model to forecast air quality

Ghanem et al. [69]

SO2,C6H6,NO,NO2,O3

Hierarchical clustering

Monitor chemicals and outline challenges related to collection and processing.

Corani [68]

Ozone, PM10

ANN, Lazy Learning

Predict levels of air pollutants from meteorological and other local variables.

Dominici et al. [67]

PM2.5

Bayesian Hierarchical Models

Assess the association of air pollution levels with the number of deaths per day

Ma et al. [58]

SO2, O3, NOx, C6H6

k-means

Developing a distributed air pollution monitoring system & use data mining to find patterns of pollutant distribution

Pegoretti et al. [62]

Rn

Geostatistical Models, KNN

Forecasting the indoor Radon concentrations

Aquilina et al. [39]

particle-associated PAH

DT, ANN

Predict personal exposure to particle-associated polycyclic aromatic hydrocarbons (PAH)

Padula et al. [57]

Traffic-related pollution

Targeted maximum likelihood estimation

Estimate the probability of low birth weight among full-term infants based on the mother’s exposure to traffic-related air pollution

Zhu et al. [35]

SO2, NO2, PM10, Respiratory diseases

ARM, GMDH

Forecasting the number of respiratory patients based on the seasonal effects of air pollution

Singh et al. [24]

AQI

PCA, Ensemble Decision DT, SVM

Predicting the Air Quality and identifying major sources of air pollution

Beckerman et al. [66]

NO2, PM2.5

GLM

Develop a better land use regression model for using machine learning methods

Pandy et al. [38]

UFP, PM

DT, RF, etc.

Test machine learning classifiers for predicting air quality and assess the impact of weather and traffic related variables on UFP and PM.

Philibert et al. [56]

N ₂0

RF

Predict NO2 emissions using variables related to chemical fertilizer treatments applied to agricultural plots.

Chen et al. [54]

Smog

ANN, Social Network Analysis

Predicting Smog based Health Hazardous regions

Dias et al. [55]

PM2.5

Density-based Clustering

Quantification of human exposure to traffic related air pollution

Lary et al. [4]

PM2.5

Ensemble Algorithms RF, SVM, ANN

Estimating the daily distributions of PM2.5

Jiang et al. [26]

AQI

Correlation Analysis

Monitoring the dynamics of air quality in large cities based on social media

Wang et al. [27]

Generic

Topic Models LDA, NLP

Evaluating the use of social media data to estimate air pollution and public response

Reid et al. [34]

PM2.5

Generalized boosting model, GAM, RF, SVM, KNN Regression, etc.

Predicting PM2.5 during wildfire

Lary et al. [52]

PM2.5

Ensemble regression models

Estimating PM2.5 distribution and relationship of such air pollutants with mental health

Lewis et al. [49]

NOx,O3, SO2, CO, VOCs, PM

Boosted regression DT, gaussian process emulation

Improve the accuracy of common low cost air pollution sensors

Hu et al. [65]

Generic

RF

Understanding, exposure to air pollution by predicting time-activity tracking of individuals

Challoner et al. [61]

PM NO2,

ANN

Predicting the indoor air quality from outdoor monitors

Mirto et al. [48]

Generic

Spatial data mining, hot spot analysis

Finding correlations between diseases and air pollution due to climatic factors

Xu et al. [30]

PM, CO O3, SO2 NO2,

SVM, Fuzzy Evaluation, Empirical Mode Decomposition

Air quality forecasting and evaluation

Min et al. [43]

PM2.5

K-Means

Apply K-Means to the identify potential new monitoring sites by considering a larger set of 313 variables in their models. Traffic and urbanicity are found to be useful to guide site selection

Keller et al. [44]

PM2.5

Modified K-Means

A clustering method to assess exposure to air pollution in health-related studies. They consider the multivariate nature of the exposure and spatial misalignment likely to occur when using data from central monitoring stations and the actual location of the cases

Liu et al. [47]

PM, SO2, CO, NO2, O3

SVM Regression

Apply support vector regression for air pollution forecasting using six criteria pollutants, five meteorological conditions and the Air Quality Index