In this project, I've used Naive Bayes implementation on several different datasets. By reporting the accuracy of the classifier, it can be found which datasets are harder to distinguish.
- How difficult it is to distinguish the difference between emails about hockey and emails about baseball?
- How hard is it to tell the difference between emails about hockey and emails about tech?
- Building an email classifier that classifies emails containing conflicting political issues (Politics_guns, Middle East, Religions)
Grading a multiple choice exam is easy. But how much do multiple choice exams tell us about what a student really knows? Dr. Dirac is administering a statistics midterm exam and wants to use Bayes’ Theorem to help him understand the following:
- Given that a student answered a question correctly, what is the probability that she really knows the material?