Package 'moc.gapbk'

Perform the Multi-Objective Clustering Algorithm Guided by a-Priori Biological Knowledge (MOC-GaPBK)

Description

This function receives two distance matrices and it performs the MOC-GaPBK.

Usage

moc.gabk(
  dmatrix1,
  dmatrix2,
  num_k,
  generation = 50,
  pop_size = 10,
  rat_cross = 0.8,
  rat_muta = 0.01,
  tour_size = 2,
  neighborhood = 0.1,
  local_search = FALSE,
  cores = 2
)

Arguments

dmatrix1	A distance matrix. It should have the same dimensions that dmatrix2. It is mandatory.
dmatrix2	A distance matrix. It should have the same dimensions that dmatrix1. It is mandatory.
num_k	The number k of groups represented by medoids in each individual. It is mandatory.
generation	Number of generations to be performed by MOC-GaPBK. By default 50.
pop_size	Size of population. By default 10.
rat_cross	Probability of crossover. By default 0.80.
rat_muta	Probability of mutation. By default 0.01.
tour_size	Size of tournament. By default 2.
neighborhood	Percentage of neighborhood. A real value between 0 and 1. It is computed as neighborhood*pop_size to determine the size of neighborhood. By default 0.10.
local_search	A boolean value indicating whether the local searches procedures (PR and PLS) are computed. By default FALSE.
cores	Number of cores to be used to compute the local searches procedures. By default 2.

Details

MOC-GaPBK is a method proposes by Parraga-Alava, J. et. al. 2018. It carries out the discovery of clusters using NSGA-II algorithm along with Path-Relinking (PR) and Pareto Local Search (PLS) as intensification and diversification strategies, respectively. The algorithm uses as objective functions two versions of the Xie-Beni validity index, i.e., a version for each distance matrix (dmatrix1, dmatrix2). More details about this compute can be found in: <https://doi.org/10.1186/s13040-018-0178-4>. MOC-GaPBK yield a set of the best clustering solutions from a multi-objective point of views.

Value

population	The population of medoids including the objective functions values and order by Pareto ranking and crowding distance values.
matrix.solutions	A matrix with results of clustering. Each column represents a clustering solution available in Pareto front.
clustering	A list containing named vectors of integers from 1:k representing the cluster to which each object is assigned.

Author(s)

Jorge Parraga-Alava, Marcio Dorn, Mario Inostroza-Ponta

References

J. Parraga-Alava, M. Dorn, M. Inostroza-Ponta (2018). A multi-objective gene clustering algorithm guided by apriori biological knowledge with intensification and diversification strategies. BioData Mining. 11(1) 1-16.

K. Deb, A. Pratap, S. Agarwal, T. Meyarivan (2002). A fast and elitist multiobjective genetic algorithm: NSGA-II. IEEE Transactions on Evolutionary Computation, 6(2) 182-197.

F. Glover (1997). Tabu Search and Adaptive Memory Programming - Advances, Applications and Challenges. Interfaces in Computer Science and Operations Research: Advances in Metaheuristics, Optimization, and Stochastic Modeling Technologies. 1-75.

J. Dubois-Lacoste, M. Lopez-Ibanez, Stutzle, T. (2015). Anytime Pareto local search. European Journal of Operational Research, 243(2) 369-385.

Examples

##Generates a data matrix of dimension 50X20

library("amap")
library("moc.gapbk")

x <- matrix(runif(50 * 20, min = -5, max = 10), nrow = 50, ncol = 20)

##Compute two distance matrices

dmatrix1<- as.matrix(amap::Dist(x, method = "euclidean"))
dmatrix2<- as.matrix(amap::Dist(x, method = "correlation"))

##Performs MOC-GaPBK with 5 cluster

example<-moc.gabk(dmatrix1, dmatrix2, 5)

example$population
example$matrix.solutions
example$clustering

---