Naive Bayes Classifier (30 — day NLP challenge)

Day 10

Naive Bayes is a probabilistic classifier based on Bayes ' theorem with the assumption of independence b/w features.

Despite its simplicity, Naive Bayes has been widely used in various applications, including spam filtering, sentiment analysis, and document classification.

Fromula :

core naive bayes

Expanded Naive Bayes Formula

Simple Example of Naive Bayes in NLP

New Sentence to Predict

We want to know: “free cash” Is this Spam or Ham?

Step 1 — Count Classes

Total sentences = 4

• Spam = 2
• Ham = 2

So:

P(Spam) = 2/4 = 0.5
P(Ham) = 2/4 = 0.5

Step 2 — Word Frequency

Spam Words

free, money, offer, win, cash, now

Ham Words

let’s, meet, tomorrow, how, are, you

Step 3 — Check Words in New Sentence

Sentence:

Both words:

• “free”
• “cash”

appear in Spam sentences.

They do not appear in Ham sentences.

Step 4 — Probability Comparison

Spam

P(Spam | free cash)
= P(free | Spam) × P(cash | Spam) × P(Spam)

High probability because both words exist in Spam data.

Ham

P(Ham | free cash)
= P(free | Ham) × P(cash | Ham) × P(Ham)

Very low probability because words are absent in Ham data.

Final Prediction

New Sentence to Predict: “free cash”

We want to know: Is this Spam or Ham?

Step 1 — Count Classes

Total sentences = 4

• Spam = 2
• Ham = 2

P(Spam) = 2/4 = 0.5 P(Ham) = 2/4 = 0.5

Step 2 — Word Frequency

Spam Words: free, money, offer, win, cash, now

Ham Words: let’s, meet, tomorrow, how, are, you

Step 3 — Check Words in New Sentence

Sentence: free cash

Both words:

• “free”
• “cash”

appear in Spam sentences. They do not appear in Ham sentences.

Step 4 — Probability Comparison

Spam:

P(Spam | free cash) = P(free | Spam) × P(cash | Spam) × P(Spam)

High probability because both words exist in Spam data.

Ham:

P(Ham | free cash) = P(free | Ham) × P(cash | Ham) × P(Ham)

Very low probability because words are absent in Ham data.

Final Prediction

Result: “free cash” → Spam

Types of Naive Bayes

1. Multinomial Naive Bayes
2. Bernoulli Naive Bayes
3. Gaussian Naive Bayes