Training Code

The training code includes the following class and functions:

class train.FuzzyArt(x, T, rho, beta, alpha, nep, update)

Train using ART Neural Network

Parameters:

• x – Input data
• rho – Free parameter
• beta – Choice Parameter
• alpha – Learning Rate
• nep – Number of epochs

create(I, T, nc, j)

Resonance did not occur - create new template

Parameters:

• I – Input
• T – Template
• nc – Number of Categories
• j – Input matrix iteration value

Return T:

Template matrix with new template

update(I, T, j, cmax)

Resonance did occur - update new template

Parameters:

• I – Input
• T – Template
• cmax – Maximum choice template location
• j – Input matrix iteration value

Return T:

Template matrix with update template

match_choice(c, norm, normI, normT)

Checks match criterion Compute choice equation Discovers best choice

Parameters:

• norm – minimum of input and templates
• normI – norm of input

Returns:

returns category choice location

template_options_loop(cmax, chmax, ch, nc)

Match Criterion

Parameters:

• cmax – Maximum choice (initialized to be -1)
• chmax – Match Criterion (initialized to be -1)
• ch – Template choice
• nc – Number of Categories

Return cmax:

Maximum choice template location

while loop end when ->

art_train(I, T, T_length)

Train ART - Create Template Matrix

Parameters:

• I – Input Matrix
• T – Template Matrix
• cmax – Max choice (initialized to be -1)
• chmax – Match Criterion (initialized to be -1)

Return T:

Final Template