Neural Network Model for Forecasting of Monthly Marine Fish Landing in Chonburi Province

Abstract

This research was aimed to build the model of simple neural network MLP for forecasting the monthly marine fish landing in Chonburi province with the number of nodes in input layer equal to all six independent variables; sea-level pressure, wind speed, total rain, rainy day, minimum temperature and maximum temperature. The root mean square error (RMSE) of validation data set was applied for measuring the performance of neural network model. The results of research revealed the MLP 6-3-1 with the hyperbolic tangent and exponential function of the activation function at the hidden and output node respectively was the best performance for forecasting the monthly marine fish landing in Chonburi province. This satisfactory result was due to the ability of this model which produced the minimum RMSE of the validation data set.