machine learning - Genetic Algorithm - Fitness function and Rule optimization -

let's have set of training examples a_i attribute , output iris-setosa

the values in dataset

a1, a2, a3, a4      outcome 3   5   2   2       iris-­setosa 3   4   2   2       iris­-setosa 2   4   2   2       iris­-setosa 3   6   2   2       iris­-setosa 2   5   3   2       iris­-setosa 3   5   2   2       iris­-setosa 3   5   2   3       iris­-setosa 4   6   2   2       iris­-setosa 3   7   2   2       iris­-setosa 

from analysis range of attribute are:

a1 ----> [2,3,4] a2 ----> [4,5,6,7] a3 ----> [2,3] a4 ----> [2,3] 

i have defined:

a1 ----> [low(2),medium(3),high(4)] a2 ----> [low(4,5),medium(6),high(7)] a3 ----> [low(<2),medium(2),high(3)] a4 ----> [low(<2),medium(2),high(3)] 

i have set below:

a1,         a2,         a3,         a4          outcome medium      low         medium      medium      iris-setosa      medium      low         medium      medium      iris-setosa low         low         medium      medium      iris-setosa medium      medium      medium      medium      iris-setosa low         low         high        medium      iris-setosa medium      low         medium      medium      iris-setosa medium      low         medium      high        iris-setosa high        medium      medium      medium      iris-setosa medium      high        medium      medium      iris-setosa 

i know have define fitness function. problem? in actual problem there 50 training examples similar problem.

how can optimize rule using ga? how can encode?

suppose if input (4,7,2,3), how optimization can me classify whether input iris-setosa or not?

thank patience.

the task describe known one-class classification.

identifying elements of specific class amongst elements, learning training set containing objects of class is

... different , more difficult traditional classification problem, tries distinguish between 2 or more classes training set containing objects classes.

a viable approach build outlier class data artificially , train using 2 class model it can tricky.

when generating artificial outlier data need wider range of possible values target data (you have ensure target data surrounded in attribute directions).

the resulting two-class training data set tends unbalanced , large.

anyway:

• if want try genetic programming one-class classification take at

  one-class genetic programming - robert curry, malcolm i. heywood (presented in eurogp'10, 13th european conference on genetic programming)

• also consider anomaly detection techniques (a simple introduction 9th week of coursera machine learning class andrew ng; notes here).