چکیده مقاله
Accurate forecasting of polyethylene granule prices plays a critical role in procurement planning, risk management, and strategic decision making within petrochemical supply chains This study presents a comprehensive machine learning based forecasting pipeline for predicting one month ahead prices of Polyethylene F7000 granules using advanced statistical and deep learning models A differencing based time series transformation combined with extensive feature engineering over 60 engineered features including lagged prices, exponential moving averages, volatility indicators, and cyclical features was implemented to enhance stationarity and generalization Multiple models including Support Vector Regression SVR , Artificial Neural Networks ANN , Long Short Term Memory LSTM , Transformer networks, and XGBoost were evaluated using rigorous train validation test splitting 70% 15% 15% Among all models, XGBoost achieved the best performance with an R² score of 97 90% and MAPE of 1 14% The findings demonstrate that tree based ensemble learning, when combined with advanced feature engineering, provides superior predictive capability for polyethylene price forecasting
کلیدواژهها
نویسندگان
شیوه ارجاع
Hamidi, Marzieh and Yasinian, Hamid and Sharifi, Yousef,1404,Machine Learning-Based Price Forecasting for Polyethylene Granules,29th National Conference on Electrical, Computer and Mechanical Engineering,Shirvan
ارائهشده در
مجموعه مقالات بیست و نهمین کنفرانس ملی مهندسی برق ،کامپیوتر و مکانیک28 بهمن 1404 · شیروان