Data Analysis of Electrical Systems Using Machine Learning Algorithms

Machine Learning for Load Forecasting in Power Systems

“…Seaborn makes it simpler to build sophisticated graphics like heatmaps, violin plots, and pair plots, to name a few. The above figures clearly represent the accuracy of different supervised Machine Learning Algorithms [10] to predict the power from generating sources like Hydro, Thermal-Steam, Thermal-Diesel, Thermal-gas, Nuclear, Renewable, and overall capacity for future purposes all over India. The objective is to predict future data associated with the "Sourcewise installed generation capacities in the country as of 30.06.2022" data file, by utilizing machine learning algorithms such as Logistic Regression, which calculates the probability for one to all by combining all decisions, giving an accuracy of 0.333, Using a data collection and precise decisions made under probabilistic settings, the Decision Tree Classifier calculates accuracy as 1.0, Combining all decisions that provide an accuracy, the Random Forest Classifier determines the chance for one to all with 0.888, Nave Bayes Classifier, which employs the Bayes Theorem to determine the likelihood of each class to produce the primary class and provides an accuracy of 0.777, Support Vector Machine, which uses hyperplanes to create functions and provides an accuracy of 0.444.…”