Unemployment rates forecasting using supervised neural networks

This study investigates the efficiency of various models used to forecast unemployment rates. The objective of the study is to find the model which most accurately predicts the unemployment rates. It starts with auto regressive models like autoregressive moving average model and smooth transition auto regressive model and then continues to explore four types of neural networks, namely multi layer perceptron, recurrent neural network, psi sigma neural network and radial basis function neural network. In addition to these, it also uses learning vector quantization in a combination with radial basis neural network. The results have shown that the combination of learning vector quantization and radial basis function neural network outperforms all the other forecasting models. It further uses ensemble techniques like support vector regression, simple average, to give even more accurate results.