naive bayes Flashcards
The “naive” in Naive Bayes refers to:
A) The simplicity of its mathematical formula
B) The assumption that all predictors are conditionally independent
C) Its use only with numeric data
D) The use of small datasets
The assumption that all predictors are conditionally independent
What does the Naive Bayes classifier estimate?
A) The linear relationship between predictors and outcome
B) The mode of the numeric predictors
C) The probability of each class given the predictor values
D) The average of categorical variables
The probability of each class given the predictor values
What is the key disadvantage of the full (exact) Bayes classifier?
A) It cannot handle categorical predictors
B) It requires calculating too many conditional probabilities
C) It requires the use of logistic regression
D) It overfits the training data by default
It requires calculating too many conditional probabilities
Why might a Naive Bayes classifier assign a class probability of zero?
A) There are missing values
B) The predictor value is continuous
C) The class does not exist in the dataset
D) The predictor value was not observed in the training set
The predictor value was not observed in the training set
What technique is commonly used to avoid zero probabilities in Naive Bayes?
A) Normalization
B) Data partitioning
C) Laplace smoothing
D) Standardization
Laplace smoothing
Which of the following is not a benefit of Naive Bayes?
A) Fast computation
B) Works well with small datasets
C) Performs well even when independence assumption is violated
D) Handles categorical variables well
Works well with small datasets
Suppose 20% of flights are delayed. A Delta flight from DCA to LGA, Sunday morning, has a 0.3 probability of delay and 0.7 on-time. According to Naive Bayes, how is it classified?
A) Delayed
B) On-time
C) Uncertain
D) Cannot be classified
On-time
What is the effect of violating the conditional independence assumption?
A) Naive Bayes fails to run
B) Naive Bayes provides better probabilities
C) Rank ordering is still preserved for classification
D) The model becomes a logistic regression
Rank ordering is still preserved for classification
Which R function is used to train a Naive Bayes model?
A) glm()
B) naiveBayes()
C) bayes.train()
D) predictNB()
naiveBayes()
What kind of predictors can the standard Naive Bayes algorithm handle directly?
A) Only numeric
B) Only binary
C) Only text-based
D) Only categorical
Only categorical
