Code for text mining through Bayes Classifier?
4 views (last 30 days)
Christopher Creutzig on 26 Nov 2018
MathWorks has examples like this one using classifiers from the Statistics and Machine Learning Toolbox to work on text data. The main problem with converting this to using a naïve Bayes algorithm is that fitcnb is not optimized to work with high-dimensional sparse data, such as a basic word count.
To alleviate that problem, use a dimensionality reduction step, such as fitlsa.