A Neural Network Approach to GSM Traffic Congestion Prediction
In this paper, the authors propose a GSM congestion prediction model based on Multi-Layer Perceptron Neural Networks (MLP-NNs) with sigmoid activation function and Levenberg-Marquardt Algorithms (LMA) using twelve month real traffic data. The trained network model was used to predict traffic congestion along a chosen route. Regression analysis between predicted traffic congestion volumes and corresponding actual traffic congestion volumes shows a correlation coefficient of 0.986. This result clearly shows the effectiveness of Artificial Neural Networks (ANN) in traffic congestion prediction and control.