note 7: gaussian discriminant analysis, qda, lda

N/A

36 Jonathan Richard Shewchuk 7 Gaussian Discriminant Analysis, including QDA and LDA GAUSSIAN DISCRIMINANT ANALYSIS Fundamental assumption: each class has a normal distribution [a Gaussian]. ! 2 1 kx µk2 X ⇠ N(µ, ) : f (x) = p exp . [µ & x = vectors; = scalar; d = dimension] ( 2⇡ )d 2 2 For each class C, suppose we know mean µC and variance 2C , yielding PDF fX|Y=C (x),…

related reading

• regressionpeople.eecs.berkeley.edu
• note 6: bayes decision rulepeople.eecs.berkeley.edu
• 02.pdfpeople.eecs.berkeley.edu
• notes.pdfjingbol.web.illinois.edu
• Six (and a half) intuitions for KL divergence — LessWronglesswrong.com
• Approximating KL Divergencejoschu.net
• 6.036_notes.pdfphillipi.github.io
• rakhlin_mathstat_sp24.pdfmit.edu
• Coordinate Ascent Mean-field Variational Inference (Univariate Gaussian Example)suzyahyah.github.io
• RescaledKL.tmdanilorezende.com
• Logs, Tails, Long Tails – Ryan Moulton's Articlesmoultano.wordpress.com
• unsupervised learning and PCApeople.eecs.berkeley.edu