The Naïve Bayes technique for predicting categorical outcomes employs both ________ and _______.

• A. Probability; Variance
• B. Probability; Mean
• C. Likelihood; Entropy
• D. Prior; Posterior

Answer: (A)

Naive Bayes classifiers rely on probabilities to decide the most likely class for a given instance. The method uses Bayes’ rule to combine how likely the observed features are under each class with how common each class is. When features are continuous, the likelihood is modeled with a probability distribution; in the common Gaussian Naive Bayes, this distribution is Gaussian and is defined by its mean and variance. The variance parameter captures how spread out the feature values are within each class and is essential to computing the likelihood P(feature|class). So, the technique hinges on probability and the variance of the feature distributions, which is why that pair best describes the components involved.