Predicting Sugar Import Quantity Using Multi-Layer Perceptron Regressor Method

This paper presents a novel approach for predicting the quantity of sugar imports using the MLP Regressor method. The study compares various Machine learning and Deep Learning techniques, including Long Short-Term Memory (LSTM), Convolutional Neural Networks (CNN), and Ensemble Methods, with the MLP Regressor model demonstrating superior performance in terms of key evaluation metrics such as Mean Squared Error (MSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and R-squared (R2) scores. The dataset’s missing values challenge is addressed by implementing the K-Nearest Neighbors (KNN) algorithm for data imputation, ensuring data integrity. The findings contribute to developing an accurate predictive model for sugar imports and emphasise the significance of handling missing values in the dataset. Further research efforts should focus on refining and enhancing the proposed model’s accuracy and reliability