Machine Learning Quiz 2 Review: KNN, Model Bias, Variance, and Classification

Vocabulary flashcards covering key machine learning concepts from Quiz 2, including k-nearest neighbors, model bias and variance tradeoffs, distance metrics, and logistic regression.

k-Nearest Neighbor (KNN) Classifier

A classifier that is often better than logistic regression for strongly nonlinear phenomenon.

Effect of Increasing $k$ in KNN

In a k-nearest neighbor predictive model, as the value of $k$ increases, the model bias increases and the model variance decreases.

Model Variance

The error that occurs when fitting spurious phenomenon in the data rather than true relationships.

Model Bias

The error that occurs when a model oversimplifies the true relationship between the outcome variable and input variables.

Model Flexibility Tradeoff

The dynamic wherein as model flexibility increases, bias tends to decrease but variance tends to increase.

Manhattan Distance

A distance metric that is distorted less by outlier observations when compared to Euclidean distance.

Logistic Regression

A classification approach that generates a linear decision boundary to classify observations.