ga_parsimony-class {GAparsimony}

Class "ga_parsimony"

Description

An S4 class for searching parsimonious models by feature selection and parameter tuning with genetic algorithms.

Objects from the Class

Objects can be created by calls to the ga_parsimony function.

Slots

call

an object of class "call" representing the matched call;

min_param

a vector of length equal to the model parameters providing the minimum of the search space;

max_param

a vector of length equal to the model parameters providing the maximum of the search space;

nParams

a value specifying the number of model parameter to be tuned;

feat_thres

proportion of selected features in the initial population. It is recommended a high percentage of selected features for the first generations;

feat_mut_thres

threshold to consider a random number between 0 and 1 is considered one if a value of the parameters-chromosome is muted. Default value is set to 0.5;

not_muted

number of the best elitists that are not muted. Default value is set to 3;

rerank_error

when a value distinct to zero is provided a second reranking process according to the model complexities is called by 'parsimonyReRank' function. Its primary objective is to select individuals with high validation cost while maintaining the robustness of a parsimonious model. This function switches the position of two models if the first one is more complex than the latter and no significant difference is found between their fitness values in terms of cost. Therefore, if the absolute difference between the validation costs are lower than 'rerank_error' they are considered similar. Default value=0.01;

nFeatures

a value specifying the number of maximum input features;

names_param

a vector with the name of the model parameters;

names_features

a vector with the name of the input features;

popSize

the population size;

iter

the actual (or final) iteration of GA search;

iter_start_rerank

iteration when ReRanking process is actived. Default=0. Sometimes is useful not to use ReRanking process in the first generations;

early_stop

the number of consecutive generations without any improvement in the best fitness value before the GA is stopped;

maxiter

the maximum number of iterations to run before the GA search is halted;

minutes_gen

elapsed time of this generation (in minutes);

minutes_total

total elapsed time (in minutes);

suggestions

a matrix of user provided solutions and included in the initial population;

population

the current (or final) population;

elitism

the number of best fitness individuals to survive at each generation;

pcrossover

the crossover probability;

pmutation

the mutation probability;

best_score

the best validation score in the whole GA process;

solution_best_score

Solution with the best validation score in the whole GA process;

fitnessval

the values of validation cost for the current (or final) population;

fitnesstst

the values of testing cost for the current (or final) population;

complexity

the values of model complexities for the current (or final) population;

summary

a matrix of summary statistics for fitness values at each iteration (along the rows);

bestSolList

a list with the best solution of all iterations;

bestfitnessVal

the validation cost of the best solution at the last iteration;

bestfitnessTst

the testing cost of the best solution at the last iteration;

bestcomplexity

the model complexity of the best solution at the last iteration;

bestsolution

the best solution at the last iteration;

history

a list with the population of all iterations;

Author(s)

Francisco Javier Martinez-de-Pison. [e-mail]. EDMANS Group. http://www.mineriadatos.com

See Also

For examples of usage see ga_parsimony.