Binary ant lion approaches for feature selection

S-Shaped vs. V-Shaped Transfer Functions for Ant Lion Optimization Algorithm in Feature Selection Problem

“…The details of the datasets including number of attributes and instances are shown in Table 1. The proposed methods are compared PSO [26], GSA [27], and two basic ALO algorithms (coded as bALO1 [6] and bALO2 [6]). Note that, bALO1 uses an s-shaped transfer function and bALO2 uses a v-shaped transfer function.…”

“…where (D) represents the classification error rate of a given classier (the K-Nearest Neighbor (KNN) classifier [28] is used here). | | is the number of selected features, | | is the total number of features in the dataset, and ∈ [1,0], = (1 − ) are two parameters corresponding to the importance of classification quality and subset length as per the recommendations in [6]. Algorithm 1 shows the pseudocode of the proposed approach.…”

“…Another work in the literature is a chaotic-based ALO algorithm [24], in which a set of chaotic functions were used to control the balance between exploration and exploitation in the original algorithm. Later on, a binary version of the ALO algorithm was proposed by Emary et al in [6]. In all the aforementioned works, two transfer functions (proposed by Kennedy and Eberhart [18] and Rashedi et al [19],) were used to convert the continuous ALO to a binary version with the eventual goal of solving feature selection problems.…”

“…In this work, we are interested to study the impacts of different transfer functions on the performance of ALO. As such, six transfer functions proposed by Mirjalili and Lewis [25] are employed and replaced by the two transfer functions used in [6]. Table 1 shows the mathematical formulation of the transfer functions used in this paper and Fig.…”

An Improved Binary Differential Evolution Algorithm for Feature Selection in Molecular Signatures

“…A multiple populations based on GA 2009 [29] Two-stage approach based on GA 2008 [30] Hybrid method of local search and GA 2004 [31] PSO with mimicking forward and backward feature selection 2014 [32] PSO with a gbest resetting mechanism by including zero features 2008 [33] Boolean particle swarm optimization with a dual-band dual-polarized printed antenna 2006 [34] PSO with an artificial immune system (AIS) called negative selection 2008 [35] A velocity bounded Boolean particle swarm optimization 2016 [36] ACO with the use of limited pheromone values 2008 [37] A discrete biogeography based optimization 2015 [38] The multiobjective biogeography based optimization 2014 [39] A binary firefly optimization algorithm 2016 [40] A binary cuckoo search algorithm 2013 [41] Bacterial algorithm with a roulette wheel weighting strategy 2016 [42] A binary ant lion approaches 2016 [43] A modified binary DE with Boolean mutation operator 2010 [46] A binary-adapted DE (BADE) 2007 [47] DE with converting real value from DE operators to binary values by Sigmoid Limiting Function 2009 [48] An adaptive DE algorithm 2013,2016 [49,50] A hybrid approach of differential evolution and artificial bee colony 2016 [51] Differential evolution with a binary mutation scheme 2016 [52] Full Paper www.molinf.com…”