Why is Naive Bayes still a solid baseline for text classification?
easyAnswer
- For bag-of-words / TF-IDF features it is (1) very fast to train and predict — a few matrix operations; (2) memory-efficient — you only store per-class word probabilities; (3) works well even with tiny labeled datasets since parameters are estimated per feature; (4) linear in the number of features; (5) resilient to irrelevant features because they contribute little to the score.
- It's the default first baseline before trying logistic regression or transformers.
Check yourself — multiple choice
- It always beats logistic regression on text
- It's fast, memory-efficient, and needs little data — a strong baseline for BoW / TF-IDF text
- It requires GPUs
- It only works on binary text tasks
Cheap, low-data, robust — NB is the go-to first baseline on BoW text.
#naive-bayes#text
Practise Supervised Learning
215 interview questions in this topic.
Related questions
- Gaussian NB vs Multinomial NB vs Bernoulli NB — when do you use each?
- How does Bayes' theorem drive Naive Bayes classification?
- The 'naive' independence assumption is almost always false. Why does Naive Bayes still work?
- What is Laplace (additive) smoothing in Naive Bayes?
- Why do we compute Naive Bayes scores in log space?
- You trained Naive Bayes on balanced data but deploy on data where positives are 1%. What happens and what do you do?