Energy Consumption Forecasting for Smart Industry Using Auto-Regressive Integrated Moving Average (ARIMA) and Vector Auto-Regression (VAR) Model

In the contemporary world, the escalating apprehensions about energy consumption's ecological repercussions have spurred urgent attention. The strategic anticipation of energy consumption trends holds pivotal importance in steering efficient energy use and curtailing the release of carbon emissions. In this comprehensive study, we proffer a novel approach, leveraging two-time series models, namely VAR (Vector Auto regression) and ARIMA (Autoregressive Integrated Moving Average), to meticulously dissect data gleaned from a smart industry context. The study's focal objective is the forecast of both energy consumption patterns and CO2 emissions, thereby illuminating the prowess of time series models in this predictive endeavor. As the crux of our investigation, we embarked on a comparative analysis of the efficacy inherent in these models. Our findings emphatically underscore the commendable predictive capacity of the ARIMA model in delineating energy consumption forecasts. Remarkably, the optimal configuration for the ARIMA model was determined as (p=1, d=1, q=0), delineating its reliance on a single antecedent value of the dependent variable and a singular differencing operation for precise forecasting. The ARIMA model further distinguished itself by manifesting a trifecta of diminutive metrics: Minimal Absolute Error (MAE), Mean Squared Error (MSE), and Root Mean Squared Error (RMSE). Such trifling metrics underscore the model's adeptness in extrapolating forthcoming energy consumption patterns with commendable accuracy. The discerning residual analysis fortified our model's credibility, as it showcased the random dispersion of errors, a testament to the model's prowess in encapsulating all pertinent data intricacies. our study underscores the instrumental role of time series models, specifically ARIMA, in proficiently prognosticating energy consumption trends. This empirical evidence serves as a clarion call for adopting these models in analogous domains to ameliorate energy consumption strategies, thereby orchestrating a harmonious accord between human progress and ecological preservation.