PM₁₀ concentration forecast using modified depth-first search and supervised learning neural network

Atmospheric particulate matter (PM) is an important factor that influences the weather and climate changes which have an impact on life and Earth. In this study, we attempt to forecast PM₁₀ (particulate matters with diameters that are less than or equal to 10 μm) concentration by using data from Nan Province of Thailand as a case study because the main agricultural occupation of Nan is corn growing and air pollution is always the major problem in this region, especially PM₁₀ that is the result from burning corn fields after harvesting. In order to forecast PM₁₀ concentration at each monitoring station 1 h ahead, a novel model based on a combination of genetic algorithm, multilayer perceptron neural network, and modified depth-first search algorithm is proposed. Experimental results show that the proposed model (in Fig. 6) performs better than other models when forecasting 1 h ahead.