Prediction of QoS Data for Pressure Sensor Using Prophet and FFNN AI Algorithms

In wireless networks, monitoring of wireless environment is performed by spectrum sensors and network sniffing. By this we will get a set of measurements which are spatially distributed. In this paper the QoS data from pressure sensor is taken and based on that, univariate models were constructed. For construction of the models, Prophet and Feed Forward Neural Network (FFNN) artificial intelligence algorithms were used. The algorithms are applied to data taken from a wireless sensor network test bed. In order to evaluate the model, we have predicted and analyzed the pressure sensor data. The RMSE of Prophet algorithm prediction is 8.6108, MASE is 0.2890 and SMAPE is 0.0961. The FFNN approach provides a RMSE of 8.0495, MASE of 0.2902 and SMAPE of 0.0967. The results indicate that FFNN supervised machine learning approach provides accurate results as compared to the prophet approach.