Parallelizing Gene Expression Programming Algorithm in Enabling Large-Scale Classification

Algorithm 1

The pseudocode of classification using GEP.

For each

    Encode representing the class of

For each in

      Input and into GEP

Let be and be .

Initiate GEP components: function set, link function, selection

    mutation, crossover and fitness .

In each generation, GEP mines ,

    if

GEP keeps running

    until terminating condition (determined generation or fitness) met

Output which represents

Let denote a threshold

    For each in

    Compute

    if

       ∈

    else

       is an outlier

Classification terminates