Machine Learning Meetup Notes: 2010-06-02

From Noisebridge
Jump to: navigation, search


  • Sampling: possible by student only, uses original data format, may need changes to work with current data. (Thomas)
  • Distribute dataset -- provide to at least Thomas
  • Superskills: Thomas & Mike attempts.
  • Naive Bayes vs LibSvm (in weka!) vs Neural networks vs Regression
  • Student classifier -- Java program to save/load models?! (can eliminate students not in test set!)
  • Format for future meetings
  • See how many students are new in the test set;
  • Create statistics/histograms for IQ/chance features -- for discretization for Naive Bayes
  • Add feature "IQ strength" & "chance strength"?