Solar Power Production Forecasting in India using Machine Learning Approach

In this work, the machine learning technique is applied for forecasting of electricity production from solar energy. Electricity consumption per capita (kWh per capita), total electricity consumption in India (GWh), and GDP per capita (INR per capita) are the most influential economic factors in solar power production in India and considered as contributory factors of the proposed model. The best model has the least error in terms of RMSE (0.28), MAD (0.20), and MSE (0.20), along with the best fit of R-squared value. The performance of proposed and conventional models is compared. The proposed model is found to be more accurate. Power generation from solar energy is forecasted 73.66 GW by 2025 and 144.66 GW by 2030. This study will help the government agencies, policymakers to frame policies accordingly. It will give a preliminary idea to the investors to make decisions precisely.