Fine-Tuning the High-Voltage Tower Pollution Model Using Distribution Identification

Modelling highly stochastic variables is often a challenging problem, as evidenced in modelling pollution on high-voltage pylons. This paper presents one approach to increasing model accuracy by identifying distributions and transforming inputs during the data preprocessing phase. The identification of statistical distributions of inputs allowed for the successful determination of the type of distribution, its parameters, and the confirmation of specific domain-specific knowledge. Notably, such transformation of input attributes enabled an increase in the accuracy of the regression model, thereby harnessing the full potential of the available data.