A Comparison of machine learning regression models for critical bus voltage and load mapping with regards to max reactive power in PV buses

• PCA is effective in generating voltage controlling areas near critical point. ANFIS and KNN were the best regression algorithms when load predictions. ANN, SVR DT showed to be inferior Voltage Loading mapping. a fast assessment tool for predictions due its rules. The aim of this paper compare system loading mapping capabilities variety algorithms, such as Adaptive Network based Fuzzy Inference System (ANFIS), Artificial Neural Networks (ANN), K-Nearest Neighbors (KNN), Support Vector Regression (SVR), Decision Tree (DT). A sensitivity matrix generated from power flow Jacobian scenario unstable Principal Component Analysis (PCA) used separate system, close point, order group buses into coherent areas. For different reactive injection scenarios, we have bus voltages that can mapped by aforementioned algorithms. are trained with limited amounts data, establish fair comparison between them. present work shows better performance prediction compared rest. academic IEEE 14 118 systems employed all limits considered, so results may reproduced.