Which of the following four modeling algorithms is least vulnerable to outlier bias? (Select one) A. Linear Regression B. Naive Bayes C. k-NN D. GLM

To determine which of the four modeling algorithms is least vulnerable to outlier bias, let's analyze each option:

A. Linear Regression: Linear regression is sensitive to outliers because it tries to minimize the sum of squared residuals. Outliers can have a large impact on the regression line, making it less robust to such anomalies.

B. Naive Bayes: Naive Bayes is generally less sensitive to outliers because it relies on the probability distribution of the features rather than their exact values. It assumes feature independence and uses the overall distribution, which can mitigate the effect of outliers.

C. k-NN (k-Nearest Neighbors): k-NN can be quite sensitive to outliers because it relies on the distance between data points. Outliers can significantly affect the distance calculations and, consequently, the classification or regression results.

D. GLM (Generalized Linear Model): GLMs, like linear regression, can also be sensitive to outliers depending on the link function and the distribution of the error terms. Outliers can influence the parameter estimates and the overall model fit.

Based on this analysis, the algorithm that is least vulnerable to outlier bias is:

B. Naive Bayes