05 Prove : Assignment
Naive Bayes Classifier
Objective
Understand the basics of the naive Bayes algorithm.
Please note that this is not a programming assignment. Instead, you will walk through the basics of the algorithm using a spreadsheet. This exercise is intentionally lighter than other weeks because you will be finishing your more involved Decision Tree assignment.
Class Example
The following videos walk through an example of this activity for a different dataset:
Naive Bayes Example - Part 1 (3 mins) - Discusses the problem to be solved and the general approach to solving it.
Naive Bayes Example - Part 2 (12 mins) - Uses a spreadsheet to calculate the probabilities for the Naive Bayes classification.
In case it is helpful, here is a link to the spreadsheet from the class example: Class Example Spreadsheet.
Instructions
Use a spreadsheet to track the probabilities of the of the following attributes. Then, using naive Bayes, make calculations by hand (in the spreadsheet), to determine the classification of the provided data points.
For this exercise, use the following simplistic data set:
|
Row # |
Credit Score |
Income |
Collateral |
Job History |
Should Loan |
|---|---|---|---|---|---|
|
1 |
Good |
High |
Good |
Short |
Yes |
|
2 |
Good |
High |
Good |
Long |
Yes |
|
3 |
Good |
High |
Poor |
Short |
No |
|
4 |
Good |
Low |
Good |
Long |
Yes |
|
5 |
Good |
Low |
Poor |
Long |
No |
|
6 |
Average |
High |
Good |
Long |
Yes |
|
7 |
Average |
Low |
Poor |
Long |
No |
|
8 |
Average |
Low |
Poor |
Short |
No |
|
9 |
Average |
High |
Poor |
Long |
Yes |
|
10 |
Average |
Low |
Good |
Long |
No |
|
11 |
Low |
High |
Good |
Long |
Yes |
|
12 |
Low |
High |
Poor |
Long |
No |
|
13 |
Low |
High |
Good |
Short |
No |
|
14 |
Low |
Low |
Poor |
Long |
No |
You should make a prediction for the following instances:
Credit Score: Good
Income: High
Collateral: Good
Job History: LongCredit Score: Average
Income: Low
Collateral: Good
Job History: Short-
Credit Score: Low
Income: High
Collateral: Poor
Job History: Short The class-conditional probabilities for each attribute.
The calculations for Yes/No classifications for each of the 3 items to predict. In other words, six formulas/results that are clearly marked.
Submission
When complete, upload a copy of your spreadsheet. Then in the "submission comments", give your prediction for the three instances above.
To receive full credit for this assignment, your spreadsheet must clearly show the following:
Please note that by the nature of this assignment, it is difficult to "show creativity" and "excel above and beyond the requirements," so there is not an expectation for that here.