Li, Fangyao

Air Pollution Prediction by Using the Matching Pursuit Algorithm

Fangyao Li¹, Christopher M. Triggs¹, Bogdan Dumitrescu², and Ciprian Doru Giurcăneanu¹

1. Department of Statistics, University of Auckland

2. Department of Automatic Control and Computers, University Politehnica of Bucharest, Romania

Particulate matter (PM) is one of the most dangerous air pollutants. PM is measured as PM_2.5 and PM₁₀, where 2.5 and 10 are the maximum diameters of the particles. PM_2.5 is much more expensive to measure compared with PM₁₀. Hence, it is important to estimate the PM_2.5 concentration for the areas where the measurements do not exist. Statistical methods are vital in prediction air pollution data. We will briefly introduce the PM data collected at four sites of Auckland by the National Institute of Water and Atmospheric Research (NIWA), then illustrate the prediction of PM_2.5 concentration by using the matching pursuit algorithm (MPA). Novel stopping rules for MPA, based on information theoretic criteria, will also be discussed (Li et al. 2018).

This presentation is eligible for the NZSA Student Prize.

References

Li, F., C.M. Triggs, B. Dumitrescu, and C.D. Giurcăneanu. 2018. “The Matching Pursuit Algorithm Revisited: A Variant for Big Data and New Stopping Rules.” Signal Processing (in press).