An insight into the estimation of relative humidity of air using artificial intelligence schemes

Abstract

The present work suggested predicting models based on machine learning algorithms including the least square support vector machine (LSSVM), artificial neural network (ANN), and adaptive network-based fuzzy inference system (ANFIS) to calculate relative humidity as function of wet bulb depression and dry bulb temperature. These models are evaluated based on several statistical analyses between the real and determined data points. Outcomes from the suggested models expressed their high abilities to determine relative humidity for various ranges of dry bulb temperatures and also wet-bulb depression. According to the determined values of MRE and MSE were 0.933 and 0.134, 2.39 and 1, 1.291 and 0.193, 0.931 and 0.132 for the RBF-ANN, MLP-ANN, ANFIS, and LSSVM models, respectively. The aforementioned predictors have interesting value for the engineers and researchers who dealing with especial topics of air conditioning and wet cooling towers systems which measure the relative humidity.