Example 1: Continuous Attribute Example

Fisher's (1936) Iris data is often used for benchmarking classification algorithms. It consists of the following continuous input attributes and a classification target:

Output

The Naive Bayes classifier incorrectly classifies 6 of the 150 training patterns.

     Iris Classification Training Error Rates
------------------------------------------------
  Setosa   Versicolour   Virginica    |   Total
  0/50         0/50        20/40       |   20/140
------------------------------------------------



Probabilities for Incorrect Classifications
 Predicted   
   Class     |  Class       |   P(0)     P(1)     P(2) 
-------------------------------------------------------
Virginica    | Virginica    |   0.000    0.436    0.564 
Virginica    | Virginica    |   0.000    0.466    0.534 
Versicolour  | Virginica    |   0.000    0.542    0.458 
Virginica    | Virginica    |   0.000    0.441    0.559 
Virginica    | Virginica    |   0.000    0.412    0.588 
Virginica    | Virginica    |   0.000    0.466    0.534 
Versicolour  | Virginica    |   0.000    0.542    0.458 
Versicolour  | Virginica    |   0.000    0.515    0.485 
Virginica    | Virginica    |   0.000    0.460    0.540 
Versicolour  | Virginica    |   0.000    0.551    0.449 
Link to Java source.