Investigations on Impact of Feature Normalization Techniques for Prediction of Hydro-Climatology Data Using Neural Network Backpropagation with Three Layer Hidden

Data normalization techniques are a very important initial stage to be carried out in order obtain good predictive data approach. Many researchers get different prediction and error results each use of these techniques. Thereby, this article discusses the accuracy rate seven at preprocessing Neural Network Backpropagation (NNBP) architecture including decimal scaling, Z-score, min-max (there 6 types), sigmoid, tanh estimators, mean-MAD, median-MAD. We used two patterns: seasonal (rainfall) stationary (air humidity) that taken over past 10 years (at 10-day intervals). parameters number epochs, MAE, MSE when conducting training, testing, predictions. The showed Z-score technique was for rainfall with epochs 10, MAE 0.051, 0.004. In case air humidity data, mean-MAD can recommended 8, 0.013, 0.0004, while 7, 0.018, 0.0006. Thus, we conclude other predict or normalization.