Cascade Forward Neural Networks-based Adaptive Model for Real-time Adaptive Learning of Stochastic Signal Power Datasets

In this work, adaptive learning of a monitored real-time stochastic phenomenon over an operational LTE broadband radio network interface is proposed using cascade forward neural (CFNN) model. The optimal architecture the model has been implemented computationally in input and hidden units by means incremental search process. Particularly, we have applied adaptive-based cascaded for realistic practical signal data taken from cellular network. performance compared with benchmark feedforward (FFNN) number measured SINR datasets obtained period three months at two indoors outdoors locations results showed that CFNN provided best (0.9310 RMSE; 0.8669 MSE; 0.5210 MAE; 0.9311 R), to FFNN (1.0566 1.1164 0.5568 0.9131 R) first studied outdoor location. Similar robust performances were attained other locations, thus indicating it superior phenomenon.