Predicting Cotton Production in Syria Using The MLP Model

2030 based on time series data on cultivated area and productivity from 1993 to 2023. The descriptive approach summarized time series data for the research variables using statistical methods like tables and charts and descriptive measures of central tendency and dispersion. A standard analytical approach was employed using a neural network model, specifically multi-layer perceptron, to fit the data. The results indicated significant instability in cotton cultivation areas in Syria, with a coefficient of variation of 59.37%. Cotton production exhibited notable fluctuations, reflected in a coefficient of variation of 68.23% while cotton crop productivity showed reasonable stability at a CV of 22.61%. The neural network model 3, 2, 1 was used in the prediction by the feed-forward method, and the backpropagation method updated the weights after calculating the difference between the expected and actual values to improve the model predictive accuracy. The neural network model proved highly efficient in predicting cotton production, recording significant decreases in the squared error 0.009 and the relative error rate 0.001, indicating its ability to analyze the data effectively. The production forecasts for the period 2024-2030 showed a positive trend with an annual growth rate ranging between 0.19% and 0.58%, indicating the possibility of achieving sustainable gains in production