Naive bayes classifier solved exercise in NLP, How to find the class of a word document using Naive Bayes classifier? Naive Bayes classifier solved example, text classification using naive bayes classifier, solved text classification problem using naive bayes
Naïve Bayes Classifier
Question:
A Naive Bayes text classifier has to
decide whether the document ‘Chennai Hyderabad’ is about India (class India) or
about England (class England).
a) Estimate the probabilities that are
needed for this decision from the following document collection using Maximum
Likelihood estimation (no smoothing).
Doc. No.

Document

Class

1

Chennai Mumbai

India

2

Delhi London Hyderabad

England

3

Chennai Kolkata

India

4

Delhi Hyderabad Pune

India

5

London Bristol Chennai

England

b) Based on the estimated
probabilities, which class does the classifier predict? Explain. Show that you
have understood the Naïve Bayes classification rule.
Solution:
a) Probability estimation
As per Naïve bayes classifier, we need two
types of probabilities namely, conditional probability denoted as P(wordclass)
and prior
probability denoted as P(class) in order to solve this problem.
Conditional
probability
Let w_{i} be a word among n words
and c_{j} be the class among m classes. The "individual"
likelihoods for every word in the word vector can be estimated via the
maximumlikelihood estimate as follows;
Here,
is the Number of times word w_{i}
appears in documents under class c_{j}
is the Count of words appears in all documents
that are listed under class c_{j}.
Prior probability
Prior probability is the total
probability of a class. That is, how often does this particular class occur in
total? This can be calculated as follows;
Here,
is the Total number of documents that are
listed under class c_{j}
is the total number of classes
For the given problem, we need to calculate
these probabilities for the test document ‘Chennai Hyderabad’. It goes as
follows;
Conditional probability estimation
P(word  class) = P(ChennaiIndia)
= 2/7
[How P(ChennaiIndia)
= 2/7? As per the training data given, only 2 documents (documents 1 and 3) are listed under the class 'India' and have the word 'Chennai'. hence, 2 in the numerator. There are totally 7 words (2 words in doc 1, 2 in doc 3, and 3 in doc 4) in all the documents under the class 'India' put together. For the remaining conditional probabilities, you do the calculation.]
P(Hyderabad  India) = 1/7
P(Chennai  England) = 1/6
P(Hyderabad  England) = 1/6
Prior probability estimation
P(India) = 3/5 [How P(India)
= 3/5? As per the training data, out of 5 documents, only 3 are listed under the class 'India'.]
P(England) = 2/5
b) To predict the correct class of the
test document ‘Chennai Hyderabad’, we need to find the posterior probability of
the test document under each class as follows;
As per Naïve Bayes, the posterior probability for n
features for a class c_{j} is calculated as follows;
P(w1, w2, …, wncj) = P(c_{j}) * P(w_{1}c_{j})
* P(w_{2}c_{j}) * … * P(w_{n}c_{j})

P(‘Chennai Hyderabad’  India) =
P(India) * P(Chennai  India) * P(Hyderabad  India)
= 3/5 * 2/7 * 1/7
= 0.6 * 0.286 * 0.143
= 0.0245
P(‘Chennai Hyderabad’  England) =
P(England) * P(Chennai  England) * P(Hyderabad  England)
= 2/5 * 1/6 * 1/6
= 0.4 * 0.167 * 0.167
= 0.0112
After the calculation, we found that P(‘Chennai
Hyderabad’  India) > P(‘Chennai Hyderabad’  England). Hence, the
predicated class of the given document is India.
***********
No comments:
Post a Comment