A common task when planning a wireless network is the analysis of propagation losses between two transceivers. Different models, from theoretical and empirical natures are proposed in the literature; however, some of them are difficult to be parametrized, and others are thought for very specific scenarios. In this work, we perform a comparison between the simplified path loss lognormal shadow fading model versus a Support Vector Machine (SVM) regressor in a WLAN network. Results show that SVMs are more accurate predicting shadow fading effects