Machine Learning Algorithms With A Precise Energy Estimation for Bangladesh

The energy estimation of a country plays a vital role in the development of a country. Estimation may deliver some significant information, as a result the national development authorities can predict the future requirements. A recent study illustrates some key facts about the estimation of energy for Bangladesh using a machine learning algorithm. However, some limitations were found such as the non-linear component of energy, input data projection, comparative study, etc. A Machine Learning Algorithm known as an Artificial Neural Network(ANN) is taken to build a desirable model to forecast the energy of Bangladesh. In addition, another machine learning technique known as linear regression is also used for comparative analysis.However, these 2 methods require some input data to estimate the output, which is addressed as the energy consumption. In this case, GDP, Population, Import and Export were taken as input data and used to estimate energy consumption. A suitable model is generated using a feed forward neural network. A nonlinear model removes the linear trend of the data.while these nonlinear data were given to the feedforward neural network for training and testing steps. In addition, future input data had been projected using the double exponential smoothing. Some measurable parameters such as MSE (Mean Square Error), MAPE (Mean Absolute Percentage Error), $\mathbf{R}^{\mathbf{2}}$ (Regression Co-efficient) were taken for each analysis and finally the best selected model was nominated to estimate and forecast the energy for Bangladesh