Salp Swarm Algorithm for Feature Selection

Introduction

• This toolbox offers a Salp Swarm Algorithm ( SSA ) method
• The Main script illustrates the example of how SSA can solve the feature selection problem using benchmark dataset.

Input

• feat : feature vector ( Instances x Features )
• label : label vector ( Instances x 1 )
• N : number of salps
• max_Iter : maximum number of iterations

Output

• sFeat : selected features
• Sf : selected feature index
• Nf : number of selected features
• curve : convergence curve

Example

% Benchmark data set 
load ionosphere.mat; 

% Set 20% data as validation set
ho = 0.2; 
% Hold-out method
HO = cvpartition(label,'HoldOut',ho);

% Parameter setting
N        = 10;
max_Iter = 100; 

% Salp Swarm Algorithm
[sFeat,Sf,Nf,curve] = jSSA(feat,label,N,max_Iter,HO);

% Plot convergence curve
plot(1:max_Iter,curve);
xlabel('Number of iterations');
ylabel('Fitness Value');
title('SSA'); grid on;

Requirement

• MATLAB 2014 or above
• Statistics and Machine Learning Toolbox