Solar Power Analysis and Estimation Using ML-Models

Estimation of solar-powered energy is becoming an essential issue in a real-time environment and friendly energy sources, and machine learning algorithms play an essential role in this area. The sunlight-based energy estimation can be viewed as standardized data. In addition, energy determination based on sunlight can be obtained from the Kaggle dataset, which assesses weather factors to focus on in the matrix. In this case, it would be helpful to know how estimation accuracy improves based on the size of the lattice hubs used in conjunction with ML algorithms. This paper used popular machine learning algorithms such as KNN, DNN, RF, and LGBM. And improve the performance tuning approach for the computation of overfitting. The ML and error calculations have shown exceptional results over time, which can be used as model data sources to predict lightning with weather conditions. From our result and discussion, we depict that LightGBM outperform other models. Random Forest and DNN demonstrate more stable and reliable predictions with fewer outliers but less compare to LGBM. The LGBM model is the best base model and outperforms other models in terms of R2 value and shows more improvement.