Forecasting Quality of Service (QoS) Parameters of IoT based Web Services with an Artificial Electric Field Algorithm trained Artificial Neural Network

Development of an accurate forecasting model for effective prediction Quality Service (QoS) parameters inter-net things (IoT) based web services is highly desired, such that it improves service management and user experience. Mostly, QoS are volatile in nature which make the IoT recommendation process chal-lenging. Artificial neural network (ANN) models found to be worthy modeling nonlinear parameter sequences. However, improper tuning ANN with conventional training algorithms may lead a suboptimal model. Nature-inspired optimization methods suitable fine have shown proficient results on real-world data mining problems. There lack need explored. We develop Electric Field Algorithm (AEFA) trained (AEFANN) where AEFA used search optimal structure. The structure achieved by through evolutionary process. Two enabled datasets evaluating effectiveness AEFANN terms three performance metrics. Experimental procedures comparative studies conducted establish superiority proposed approach over four other similar forecasts. obtained relative worth values 4.13% ~ 69.12 % (5-min granularity) 43.32% 80.3 (1-hr from SERVICE 1 dataset. Similarly, 7.25% 65.57 43.38% 72.43 2 dataset when compared oth-er models. This significant improvement existing