Particulate Matter Forecasting in Border Region Polish and Czech Republic by using Data Mining Methods
The paper presents an attempt to analyze the influence of meteorological factors on the emergency of episodes of high concentrations of particulate matter in border region Polish and Czech Republic for common city Cieszyn-Cesky Tesin. The study analyzed the air pollution principal causes and identified the best subset of features: meteorological data and air pollutants concentration, in order to predict its short-term concentration. The data comes from data acquisition stations belongs to Automatic Monitoring Air Quality and installed in the both Cieszyn and Cesky Tesin. The system analyzes the data using data mining as artificial neural networks, including self-organizing Kohenen network with exhaustion and fuzzy sets. The system creates the forecast using Case Base Reason technology. In addition to the data acquisition stations, data is input using the results of the numeric weather forecast. The idea of forecasting in general consist of two phases: classification of meteorological situations responsible for high concentration of air pollution and calculation of daily runs (one hour step) of air pollutants concentration for selected set of meteorological parameters responsible for hourly concentrations of pollutants within a defined range of values. The forecast includes hourly calculated prediction values of the concentration PM10 and also a daily contaminations factors called AQI (Air Quality Index, EPA, US).The final aim of the research have the implementation of a prognostic tool able to reduce the risk for the particulate matter concentrations to be above the alarm thresholds fixed by the law.