My task is 'unemployment forecasting'. I have data with monthly ammount of unemployed. First as I do was : simple exp smoothing, holt's method, Holt-Winters method, ARMA model. All this models take only 1 series and build prediction based only on historical data.
Now I need to forecast unemployment using such data as unemployment benefits, gross domestic product, Consumer price index, etc.
My scientific adviser suggest to use neural network. But I dont think that it will do good forecasts (because of small amount of features).
What methods (which based not only on 1 column, but on all predictors) can I use for forecasting?