K-means Clustering and Regression Analysis Study Guide

This set of flashcards covers key terms and concepts related to K-means clustering, hierarchical clustering methods, unobserved heterogeneity, fixed effects, random effects, and nonlinear regression from the study guide.

K-means clustering

A method that partitions numeric data into k compact clusters by minimizing within-cluster variance.

Hierarchical clustering

A method that builds a dendrogram to organize clusters, favoring compact groups with Ward linkage.

Unobserved Heterogeneity (UH)

Differences in data units that are not fully observed, impacting outcomes and potentially biasing ordinary least squares (OLS) regression.

Fixed effects (FE)

A model that removes time-invariant traits by focusing on within-unit changes.

Random effects (RE)

A model that treats unit traits as random, under independence assumptions.

Quadratic regression

A nonlinear regression model that includes a squared term to capture curvature in the relationship between variables.

Ward's Linkage

A method in hierarchical clustering that minimizes within-cluster variance.

Single Linkage

A hierarchical clustering method that measures the distance between clusters as the distance between their closest points.

Complete Linkage

A hierarchical clustering method using the distance between the farthest pair of points in clusters.

Average Linkage

A hierarchical clustering method using the average distance between all pairs of points in two clusters.

Convex relationship

A relationship where the second derivative is positive, indicating diminishing returns or a U-shaped curve.

Concave relationship

A relationship where the second derivative is negative, indicating an inverted-U shape.