Scribd
Upload
2 views1 page

Some Exercise Problems

The document outlines three distinct applications of probabilistic models: medical diagnosis using two symptoms, email classification into three categories, and fraud detection with three features. It presents specific probabilities for diseases, email content, and fraudulent transactions, along with the need to calculate posterior probabilities and classifications based on given data. Each section requires the application of Bayes' theorem and the assumption of conditional independence.

Uploaded by

chocomilkshake666
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
2 views1 page

Some Exercise Problems

The document outlines three distinct applications of probabilistic models: medical diagnosis using two symptoms, email classification into three categories, and fraud detection with three features. It presents specific probabilities for diseases, email content, and fraudulent transactions, along with the need to calculate posterior probabilities and classifications based on given data. Each section requires the application of Bayes' theorem and the assumption of conditional independence.

Uploaded by

chocomilkshake666
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

Medical Diagnosis with Two Features

1. A disease prediction model uses two symptoms 𝑆1 and 𝑆2 .


𝑃(𝐷) = 0.1 and 𝑃(𝐷𝑐 ) = 0.9
𝑃(𝑆1 |𝐷) = 0.8 and 𝑃(𝑆2 |𝐷) = 0.7
𝑃(𝑆1 |𝐷𝑐 ) = 0.2 and 𝑃(𝑆2 |𝐷𝑐 ) = 0.1
Assuming conditional independence, calculate: 𝑃(𝐷 ∣ 𝑆1 , 𝑆2 ).
Advanced Spam Classification
2. An email classifier has three classes:
S: Spam
P: Promotion
N: Normal
The priors are:
𝑃(𝑆) = 0.40, 𝑃(𝑃) = 0.35, 𝑃(𝑁) = 0.25
For an email containing the words 𝑤1, 𝑤2 , 𝑤3 :
|𝐶𝑙𝑎𝑠𝑠|𝑃(𝑤1 ∣ 𝐶)|𝑃(𝑤2 ∣ 𝐶)|𝑃(𝑤3 ∣ 𝐶)|
|---|---:|---:|
| Spam | 0.8 | 0.7 | 0.6 |
| Promotion | 0.6 | 0.8 | 0.5 |
| Normal | 0.1 | 0.2 | 0.1 |
Assuming conditional independence, classify the email.
Fraud Detection with Three Features

3. A bank has:
𝑃(𝐹𝑟𝑎𝑢𝑑) = 0.02 and 𝑃(𝐿𝑒𝑔𝑖𝑡𝑖𝑚𝑎𝑡𝑒) = 0.98
For a transaction:
|Feature|𝑃(𝐹𝑖 ∣ 𝐹𝑟𝑎𝑢𝑑)|𝑃(𝐹𝑖 ∣ 𝐿𝑒𝑔𝑖𝑡𝑖𝑚𝑎𝑡𝑒)|
|---|---:|---:|
| International transaction | 0.80 | 0.15 |
| High amount | 0.70 | 0.10 |
| Unusual location | 0.60 | 0.05 |
Calculate the probability that the transaction is fraudulent.

4. Suppose: 𝑃(𝐶1 ) = 0.30, 𝑃(𝐶2 ) = 0.70


For an observed feature vector X:
𝑃(𝑋 ∣ 𝐶1) = 0.8
𝑃(𝑋 ∣ 𝐶2) = 0.2
Calculate the posterior probabilities.

You might also like