Exploratory Data Analysis and Comparison of Total Energy Consumption in Major World Countries Using Artificial Intelligence

Abstract

Accurate power forecasting for electrical  consumption prediction is crucial for national energy  planning since it is a method for understanding and  anticipating market energy demand. In a deregulated  market, their power output can be modified accordingly. In  order to explicitly deal with seasonality as a class of time 

series forecasting models, Persistence Models (Naive  Models), Seasonal Autoregressive Integrated Moving  Averages with Exogenous regressors, and Univariate  Long-Short Term Memory Neural Network (LSTM) are  utilized. The main goal of this project is to do an  exploratory data analysis of the major nations' power  systems, followed by the application of several forecasting  models to determine the daily peak demand and the next  24 hours' worth of energy need once every day. Energy is  a vital resource for society's development. The  development of the economy depends heavily on an  accurate assessment of its consumption. A database of  historical energy consumption statistics from all the main  nations of the globe was created. The complete information  was modelled and simulated using machine learning  techniques to anticipate the energy consumption of each  country because national patterns may be transferrable  from one country to another. Understanding the commonalities between the characteristics of other nations  might boost prediction accuracy and enhance the  corresponding public policies.