Prediction of density limit disruption boundaries from diagnostic signals using neural networks