In machine learning, is it always true that you will achieve higher classification accuracy if you use more features ? In other words, does more features always mean higher accuracy ?
This is the question that I'm going to analyze and answer in my blog post: http://heraqi.blogspot.com.eg/2015/11/is-more-features-the-better.html. I hope you find it useful. Please let me know if you have any questions in the comments, I will be happy to answer.
For that we will use the Skin Segmentation Dataset (Rajen Bhatt, Abhinav Dhall, UCI Machine Learning Repository) and the Naive Bayesian Classifier. I implement the Bayesian classifier, and no libraries or Toolboxes are used.
Hesham Eraqi (2021). Bayesian Classifier (How many features is best ?) (https://www.mathworks.com/matlabcentral/fileexchange/58742-bayesian-classifier-how-many-features-is-best), MATLAB Central File Exchange. Retrieved .
The smallest set of features is best. For example in BMDS, tumbling re-entry vehicles generate different RCS scintillation rates, discriminating between decoys and MIRVs. RCS and velocity features are only marginal discriminants.
Find the treasures in MATLAB Central and discover how the community can help you!Start Hunting!