Comparison between ANFIS and ANN for estimation of the thermal conductivity coeffcients of construction materials

Determination of the thermal conductivity coecient of construction materials is very important in terms of fullling the condition of comfort, durability of construction materials, and the economy of country and individual. In this study, linear regression, Adaptive Neural based Fuzzy Inference System (ANFIS), and Articial Neural Networks (ANN) models were developed to estimate the thermal conductivity coecient values from the surface density (dry specic gravity/thickness) and unit weight of construction materials. Validations of the developed models were investigated by statistical analyses. In the predictive models, while the lowest determination coecient (R2) and the highest Root Mean Square Error (RMSE) were obtained from linear regression, the highest R2 and lowest RMSE were obtained from the ANFIS model. Results of the ANN model, according to the results of linear regression, showed that while R2 increased by approximately 6%, RMSE decreased by 30-39%. The results of ANFIS model revealed that while R2 increased by approximately 12%, RMSE decreased by 59-71%. As a result, it is suggested to be, along with surface density and unit weight with ANFIS which are the most appropriate methods between the used methods, an alternative approach to estimate the value of thermal conductivity