Building maintenance costs analysis in a Brazilian federal university using the simple exponential smoothing model (SES)

Abstract

At the UFMS there are around 280 Service Orders, from 2018 to 2022, issued with building maintenance services carried out in the University City, which were classified and quantified in an attempt to establish a prognosis based on existing statistical models. The objective of this work was to verify if the Simple Exponential Smoothing (SES) model can be used as a statistical tool to analyze building maintenance costs in higher education institutions, the presence of missings and possible outliers. A time series decomposition was performed using the RStudio computational tool to obtain the graphs with which the associated trend, seasonality and error can be assessed. As a main result it is possible to forecast maintenance costs using the SES model, so that the amount of costs forecast is almost equal to the amount of costs effectively demanded