Multifaktorové hodnocení rizik ve vztahu k výstupům metanu

Abstract

The aim of this thesis is to apply and describe various methods of assessing the spatial risk of methane emanation. Methods of multicriterial evaluation and machine learning and previous results of application of these methods for methane risk assessment are discussed. In the practical part of the thesis, predictive models based on support vector machines and gradient boosting machines are created and described. Based on the comparison of success of individual models and sensitivity analysis, recommendations for the use of these methods are described.