clalexni / naive-bayes

naive bayes classification

Goal

• Learn from data using naive bayes classifier
• output learn result and accuracy using test set

Requirements

• install Python 3.6
• data file (path) for inputs training and test set
• command line:
  • run these two lines to ensure the correct python version
  • the second line takes two input of data file path (./data/train.dat for instance)
    • the first input is the training set and the second input should be the test set

chmod u+x naive_bayes.py

./naive_bayes.py [input1] [input2]

• for instance:

./naive_bayes.py data/train.dat data/test.dat

• use the following command to redirect output, for example:

./naive_bayes.py data/train.dat data/test.dat > out.txt

• tested on macos/Linux

Personal Notes:

• what is naive bayes?
  • use conditional independence assumption to do classification
• what to keep track of?
  • count instances of each class value in col_values[-1]
  • count instances of attribute value given a class value in col_values[i] from i = 0 to -2
• what to do after that?
  • translate count to probability or cond prob
  • implement argmax