Classification with Naïve Bayes
ABSTRACT
The Naive Bayes series of algorithms are some of the simplest classification algorithms, but they tend to offer reasonably good results very quickly for a number of problems, including Natural Language Processing problems such as spam classification, as well as more classical feature-driven classification. In this talk, we will look at the math behind Naive Bayes classification, solving problems by hand before looking at a package in R which solves the problem for us. By the end of this talk, you should be able to apply Naive Bayes to existing problems. No experience with statistics is required, although there will be a small amount of math.ADDITIONAL MEDIA
No recordings or additional media are available for this talk.SLIDES
Click here to access the slides for this presentation.
The slides are licensed under Creative Commons Attribution-ShareAlike.
DEMO CODE
Click here to access demo code for this presentation.
The source code is licensed under the terms offered by the GPL.
LINKS & FURTHER INFO
- The golf example I walk through comes from the Naive Bayes explanation at Geeks for Geeks.
- Bruno Stecanella's practical explanation of a Naive Bayes classifier inspired my baseball versus business example and includes a good example of Laplace smoothing.
- The movie review notebook comes from Rohit Katti's R notebook on movie reviews.
- Sunil Ray has a straightforward explanation of Naive Bayes.
- Jason Brownlee has some tips and tricks for using Naive Bayes in machine learning exercises.
- Riaz Khan covers some of the math of Naive Bayes classifiers that I glossed over.
- Madhur Modi, et al, use the e1071 package to perform Naive Bayes classification against the Titanic data set.
- The naivebayes R package (PDF warning) is a great package for building Naive Bayes models.
- There is an argument that Naive Bayes isn't really Bayesian. I don't cover that at all in the talk, but thought it was interesting enough to share.