The document discusses using regression models and k-means clustering to improve voltage stability in power systems. It begins by introducing the concepts of voltage stability and issues that can cause instability. Regression models are then presented as a way to model the relationship between pre-disturbance operating points and critical voltage stability margins. The k-means clustering algorithm is also described as a method to group inputs and reduce dimensionality for improved generalization. Results show that regression models can approximate stability margins and k-means clustering effectively handles large amounts of data by determining an optimal number of cluster centers. The proposed approach is concluded to analyze the most critical post-disturbance stability margin through regression and clustering techniques.