Application of hybrid models based on deep machine learning in smart agriculture (Case studies of forecasting the future price of subtropical fish (Scomberomorus guttatus) and subtropical fish (carps and salmo trutta))

Document Type : scientific research article

Authors

1 Faculty member of Agricultural Economics Planning and Rural Development Research Institute and Head of the Secretariat of the Biodiversity Convention of the Research Organization

2 Researcher of the Planning Research Institute, Agricultural Economy and Rural Development

10.22069/japu.2024.22155.1850

Abstract

Forecasting the price and the trend of its changes is one of the most important factors in making decisions and formulating strategies related to agricultural products. The aim of the present study was to present a model with a hybrid data mining pattern including a set of nonlinear models, wavelet transform, deep neural network and Monte Carlo method for accurate forecasting of fishery products. In this research, the wavelet function is used to denoize the price data, the deep neural network is used to predict the price, the Monte Carlo method is used to simulate the most likely probability of the price, and finally, soft complex calculations are used to perform "out-of-sample prediction with new data sets". It was used for the period of 2013 to 2014. To predict the future price of tropical and cold water fish, the number of optimal breaks affecting the price is determined. In this study, a proposed hybrid data mining model was used to find the best model for predicting the future dynamic price of tropical and cold water fish. and Monte Carlo simulation has been used to increase the prediction accuracy, and the prediction accuracy of the proposed model was evaluated with three competing models using the evaluation criteria of root mean square error (RMSE) and coefficient of determination (R2). Also, the comparison of prediction accuracy The proposed model with other models shows that the use of deep neural network, compared to simple neural network, increases the power of predicting the price of agricultural products (the future price of fish). The final result shows that the use of wavelet function and Monte Carlo based on the network Deep neural network increases the accuracy of forecasting the price of agricultural products.

Keywords

Main Subjects

Journal of Utilization and Cultivation of Aquatics
Volume 14, Issue 1
April 2025
Pages 91-103

Files

Share

How to cite

Statistics