I. Introduction
Air pollutants are substances that harm the human health and the environment. They can be released from natural and anthropogenic sources: mining activities, agriculture, waste treatment, industrial processes, energy plants, burning fossil fuels, road transportation, residential activities, and natural phenomena. Air pollution was the fifth global risk factor in 2017 [1], by total number of deaths from all causes, ages, and both sexes and causes around 400 000 premature deaths per year in Europe [2]. There is also a negative economic impact by reducing productivity through working days lost, increasing health care costs, and human life span shortening. The most dangerous pollutants in Europe are particulate matter (PM), nitrogen oxides (NOx), and ground-level ozone (). Air pollution forecasting is a prevention measure against the short-term (hours or days) and long-term (years) exposure to harmful substances for vegetation, animals, and humans. With accurate forecasting, governments and policy-makers could raise real-time “low air-quality” alerts. With timely warnings, the public minimizes negative health effects by reducing outdoor activities during dangerous times. Our objective is to find the most accurate, country-scale, next-day ozone forecasting model possible.