Study of The Ann Model Performance Criteria For The Prediction of Time Series Humidity
The objective of this study is to develop a mathematicalmodel based on the Multilayer Perceptron (MLP) Artificial Neural Networks (ANN) to predict meteorological parameters in general and moisture in Particular. For this purpose, we used a time series of moisture, Measured in the area of Chefchaouen in Morocco, which depends on the air temperature, dew point temperature, atmospheric pressure, visibility, cloud cover, wind speed and precipitation. Furthermore, to choose the best architecture of the MLP neural network, we used several statistical Criteria such as: Root Mean Squared Error (RMSE), Mean Absolute Percentage Error (MAPE), Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), Mean Absolute Error (MAE) and correlation coefficient (R). The obtained results of the MLP artificial neural network are discussed and compared to the Multiple Linear Regression (MLR) traditional method. Consequently, MLP method presents a very powerful ability to predict relative moisture.
Keywords- Artificial Neural Networks, Criteria Information, Moisture spleen, Multiple Linear Regression), Prediction.