“…Surrogate models can be based on different modeling approaches, such as: polynomial response surfaces; support vector machines; space mapping; kriging; ANNs (Güneş et al, 2007(Güneş et al, , 2014De Tommasi et al, 2010, 2011Yelten et al, 2012). The ANN computational approach has gained recognition as an unconventional and a very useful modeling tool in the area of microwaves (Zhang and Gupta, 2000;Rayas-Sanchez, 2004;Marinković et al, 2008Marinković et al, , 2010Marinković et al, , 2012Marinković et al, , 2013Kabir et al, 2010;Agatonović et al, 2012;Hayati and Akhlaghi, 2013). A feature of neural networks, qualifying them to be used in various modeling applications, is their ability to learn the dependence between two data sets.…”