Variable Selection in Count Data Regression Model based on Firefly Algorithm

Abstract

Variable selection is a very helpful procedure for improving computational speed and prediction accuracy by identifying the most important variables that related to the response variable. Count data regression modeling has received much attention in several science fields in which the Poisson and negative binomial regression models are the most basic models. Firefly algorithm is one of the recently efficient proposed nature-inspired algorithms that can efficiently be employed for variable selection. In this work, firefly algorithm is proposed to perform variable selection for count data regression models. Extensive simulation studies and two real data applications are conducted to evaluate the performance of the proposed method in terms of prediction accuracy and variable selection criteria. Further, its performance is compared with other methods. The results proved the efficiency of our proposed methods and it outperforms other popular methods.

Sparse Poisson regression via mixed-integer optimization

“…We use the Python scipy.optimize module (meth-od='SLSQP') to solve the NLO problem by Eqs (20)- (22). We use Gurobi Optimizer 8.1.1 (https://www.gurobi.com/) to solve the MIQO problem by Eqs ( 13)- (17), and the indicator constraint to impose the logical implication of Eq (15). We fix the L 2 -regularization parameter to α = 0 in Tables 1, 2 and 6, whereas we tune it through hold-out validation using the training instances in Tables 3, 4 and 7.…”

“…In contrast, stepwise selection [15,16], which repeats addition and elimination of one explanatory variable at a time, is a practical method for sparse estimation. Several metaheuristic algorithms have been applied to subset selection for Poisson regression [17,18], and various regularization methods have been recently proposed for sparse Poisson regression [19][20][21][22]. Note, however, that these (non-exhaustive) sparse estimation methods are heuristic algorithms, which cannot verify optimality of an obtained subset of explanatory variables (e.g., in the maximum likelihood sense).…”

Active Effects Selection which Considers Heredity Principle in Multi-Factor Experiment Data Analysis

“…A genetic algorithm (GA) would be utilized as the optimization technique in order to reach the desired result. The use of this algorithm is similar to other metaheuristic algorithms such as firefly algorithm implemented by Algamal [2].…”