Charlene de C. Silva, Maria R. Madruga, Héliton R. Tavares, Terezinha F. de Oliveira, and Augusto e C.F. Saraiva
Beta regression, neutralization index, power equipment, insulatingoil
The normal linear regression model is usually used in different application areas of statistical techniques to investigate the relationship between a variable of interest and a set of explanatory variables, which affect the behaviour of the variable of interest through a regression structure. However, when this variable assumes values in a (0,1) interval, the use of this model is not suitable, since it assumes that the data are normally distributed. This results in a flawed model, because it enables the prediction of values outside the support of the response variable. In this context, this paper presents the use of the beta regression in the data modelling of the physicochemical and gas chromatographic characteristics used to monitor the shelf life of power equipment. The variables colour, dielectric rigidity, water content, interfacial tension, O2, N2, CH4, CO and C2H4 were statistically important in explaining the behaviour of the response variable neutralization index. Results demonstrated that the neutralization index is more likely to increase (19.73%) with the positive variation of one unit of the variable colour, and that the chance of the index increasing decreases (−7.65%) with the increase of one unit of the variable interfacial tension.
Important Links:
Go Back
