Modeling and forecasting bitumen sales using univariate and multivariate time series approaches: a case study of Algeria's NAFTAL corporation
Domaine:
environment and energyCréateur:
YouMouMohMic
Éditeur:
Ama
Hôte:
This study models and forecasts the NAFTAL Corporation’s bitumen sales in Algeria using univariate and multivariate time series methods. This method uses the Box-Jenkins method to identify, estimate, diagnostic and predict. ANOVA and ADF tests showed seasonal patterns and stationary data. The best univariate models for bitumen (BTM) and pure bitumen (BTMP) were found to be SARIMAX(2,2,1) and SARIMAX(0,1,4), respectively. No volatility clustering or ARCH effects were captured via ARCH-LM tests. VAR(2) model used to explain dynamic interactions over time. Granger causality tests detected a significant unidirectional effect between BTM and BTMP (p < 0.001 at lag 2). According to the forecast error variance decomposition (FEVD), BTM accounts for 29.44% of the 12-month forecast error variance for BTMP and 98.25% of its own variation. These findings emphasize how crucial it is to take into account the overall bitumen trends when estimating product demand.
Similaires
A Comparison of Univariate and Multivariate Time Series Approaches to Modeling Currency Exchange RateModeling and Forecasting Inflation in Ethiopia Using Multivariate Time Series AnalysisForecasting Dengue Incidences in Bangladesh: A Univariate Time Series ApproachTIME SERIES CLUSTERING IN MULTIVARIATE PRICES MODELINGSARIMA Time Series Modeling for Forecasting Malaria Positivity Rate: A Case Study of Kisumu County, KenyaHodal66/Crop-Price-Forecasting-in-Rwanda-using-Time-Series-Modeling