Forecasting with incomplete set of factors determining the predicted factor. Neural network error extrapolation method
А. В. Гусев, Alexander Okunev · International Journal of Applied Mathematics & Statistics/International journal of applied mathematics and statistics · 2017
Forecasting method that works with incomplete set of factors determining predicted factor is described in this article. The method is based on use of neural network error as a scalar value of impact of unknown factors vector on predicted factor. Authors make a conclusion that in order to forecast with required accuracy it’s necessary to extrapolate scalar value (network error).