Front. Appl. Math. Stat. Frontiers in Applied Mathematics and Statistics Front. Appl. Math. Stat. 2297-4687 Frontiers Media S.A. 10.3389/fams.2026.1744058 Original Research Improving parameters estimation of a truncated Poisson regression model based on meta-heuristic optimization algorithms Basheer Ghalya Tawfeeq 1 Conceptualization Formal analysis Software Writing – original draft Writing – review & editing Waleed Mahmood Shaimaa 2 Conceptualization Formal analysis Methodology Resources Writing – original draft Writing – review & editing Algamal Zakariya Yahya 2 * † Supervision Validation Writing – review & editing Department of Operations Research and Intelligent Techniques, University of Mosul, Mosul, Iraq Department of Statistics and Informatics, University of Mosul, Mosul, Iraq Correspondence: Zakariya Yahya Algamal, zakariya.algamal@uomosul.edu.iq

ORCID: Zakariya Yahya Algamal orcid.org/0000-0002-0229-7958

 04 02 2026 2026 12 1744058 11 11 2025 09 01 2026 12 01 2026 Copyright © 2026 Basheer, Waleed Mahmood and Algamal. 2026 Basheer, Waleed Mahmood and Algamal https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

The paper discusses computational and numerical challenges that are associated with the truncation of the information and which change the usual Poisson likelihood by the introduction of black kite optimization algorithm. Real data is used to demonstrate a significant improvement in healthcare and medical research. Truncated Poisson regression models (TPRM) are essential for analyzing count data where zero counts are unobserved, a common scenario in many real-world applications, such as healthcare and medical research. However, parameter estimation in such models often suffers from bias and inefficiency due to the complexity induced by truncation. This study proposes an improved parameter estimation approach for TPRM by leveraging meta-heuristic optimization algorithms. Specifically, we integrate state-of-the-art meta-heuristics, black kite optimization algorithm (BKA), to optimize the likelihood function and overcome the limitations of traditional iterative methods such as Newton-Raphson and quasi-Newton algorithms. Using extensive Monte Carlo simulations and real data application, we evaluate the performance of the proposed method under varying sample sizes and covariate structures. The results demonstrate that our meta-heuristic-based estimator significantly reduces mean squared error compared to conventional estimators, enhancing model reliability and predictive accuracy. The proposed approach offers a robust and efficient alternative for parameter estimation in truncated Poisson regression, with potential applications in epidemiology, ecology, and other fields dealing with truncated count data.

 BFGS black kite optimization algorithm count data meta-heuristic optimization algorithms truncated Poisson regression model The author(s) declared that financial support was not received for this work and/or its publication. section-at-acceptance Statistics and Probability Introduction

Count data modeling is a statistical field of study directed at such variables as dependent variables that are identifications of a non-negative integer number of events or occurrences, i.e., the quantity of visits to a doctor, traffic incidents, or children within a family [1, 2]. Leading to the right skewed distribution, among other characteristics unlike continuous outcomes of count data are discrete that tend to have a distorted distribution centered on lower values with a high proportion of zeros [3].

Poisson regression, a classical modeling strategy of the count data, presupposes the idea of the counts being distributed according to Poisson distribution with its means (equidispersion). This model presents the expected count as an exponential of explanatory variables that makes its interpretation to be simple in the generalized linear model framework. Empirical data however, generally do not fit the equidispersion assumption, therefore assuming they are over dispersed, which is very common, or under dispersed, which is rare. Such cases are dealt with by substituting such models with newer models with more parameters such as the Negative Binomial regression that has an extra parameter to fit flexibility to variance independent of the mean. Other distributions intended to handle different dispersion are the Generalized Poisson Conway-Maxwell-Poisson (COM-Poisson) and similar distributions [4, 5].

The other difficulty which is commonly seen is the occurrence of excess zeros; that is more counts than are predicted by standard count models. At the cost of this, zero-inflated models and hurdle models have been designed. The two components have independent structure each component modeling the result on the binary zero-non zero and the count distribution of the positive values, and thus generates better fitting and inference in situations where zeros are produced by separate mechanisms [6, 7].

The truncated Poisson regression model is a statistical technique used to analyze a count-type of data that are truncated; such as dynamic observations are excluded in the sample when smaller or larger than some integer limits [8, 9]. Truncated Poisson regression contrasts with the usual Poisson regression model in that the counts take a truncated distribution due to the requirement that the counts exceed (left truncation) or are less than (right truncation) a given value [10]. This method would consider the fact that the realized counts do not fall into the grand scope of counts that might be observed hence avoid bias in parameter estimation and wrong inference that would have been made had truncation not been considered [11, 12].

Optimization means achieving the best result under certain conditions. Engineers must make many technical and managerial decisions throughout the design, construction, and upkeep of any engineering system. The main purpose of all these decisions is to either minimize the effort or to maximize the benefit. In any situation, the effort required or the amount of benefit can be defined as a function of certain decision variables. Optimization is then the process of choosing the conditions that result in the highest or lowest value of a function [13, 14].

There is no one way that can solve all optimization problems efficiently. Due to this, various optimization methods have been invented to handle many types of optimization problems. These methods are also called mathematical programming and are typically included in the field of operations research. Operations research uses scientific methods to solve problems related to decision making and to find the best solutions [13, 14]. In recent years, new optimization methods have gained popularity and are widely used to solve difficult engineering optimization problems. Such as genetic algorithms, simulated annealing, particle swarm optimization, ant colony optimization, black-winged kite optimization algorithm [13].

The black kite optimization algorithm (BKA) is a meta-heuristic optimization algorithm based on the migration and hunting patterns of the black kite. The BKA brings together the Cauchy mutation technique and the Leader strategy to increase its ability to find global solutions and speed up convergence. This method manages achieves a balance both global and local knowledge to find good solutions [15]. It should be noted that not all problems can be solved by one algorithm. There is no meta-heuristic algorithm that is optimal in all optimization problems. A meta-heuristic algorithm can do very well with certain types of issues, but may not be as successful with others. The advancement of technology and the rise in problem complexity make it hard for some traditional algorithms to solve them [15].

The key shortcoming of parameter estimation in the truncated Poisson regression models can be viewed as the complex effect of the likelihood that is induced by truncation as it may necessitate the employment of the non-linear optimization tools and cause other problems with acquiring consistent and efficient estimators. Truncated Poisson models can be estimated using maximum likelihood in which the likelihood functions are complex with the normalizing constant being dependent upon the truncation limits. That increases the computational burden of parameter estimation and frequently requires non-linear optimization techniques iterative in nature, which have the potential to converge slowly or at local as opposed to global optima [16].

The weaknesses of quasi-Newton algorithms in parameter estimation of the truncated Poisson regression model mostly follow these issues because of the difficulty of the truncated likelihood function, which influence convergence and speed of computation. In the truncated Poisson measuring and computing log-likelihood, the log-likelihood contains a normalizing constant that relies on the truncation that results in the likelihood function to be non-linear and consequently more strenuous compared with that in the conventional Poisson. Quasi-Newton optimization methods do not compute the Hessian explicitly but can become sensitive to this complexity and thus slow to converge or to find global (as opposed to local) maxima. Quasi-Newton optimization can be sensitive to starting parameter values, and is iterative, whereas truncation or small sample size limits often render it insensitive to such start values. Such sensitivity may induce instability or convergence to irreliable parameter estimates [17, 18].

The paper therefore concerns the difficulty of causing the parameters of the truncated Poisson regressions to get the accurate estimation of parameters based on maximum likelihood approach algorithms. In particular, the paper discusses computational and numerical challenges that are associated with the truncation of the information and which change the usual Poisson likelihood by the introduction of black kite optimization algorithm. The complexity typically gets in the way of rapid convergence, arithmetical precariousness and can be biased when using traditional optimization schemes like Newton-Raphson, Gauss-Newton, or perhaps quasi-Newton procedures.

 Poisson regression

The modeling of count data is based on Poisson regression. It was the earliest model to explicitly model counts and it remains at the foundation of the numerous kinds of count models that can be used by analysts, it is also called log-linear regression. The dependent variable in this model is Poisson distributed and the dependent variable is linked with the independent variables using a log link function which yields a linear equation [4, 6].

The Poisson regression model states that the yi is a random variable distributed as Poisson with parameter λi, which depends on the regressors xi. The main formula of the model is [19]:

P(Yi=yi|xi )=e-λλyiyi!,  yi=0,1,2…,

The most popular formulation for λi is the loglinear model

lnλi=xiTβ

where xiTβ=β0+β1xi1+…+ βnxin, xiT represents the vector of independent variables and β is the regression coefficients. The expected number of events is given by

E[yi|xi ]=Var[yi|xi ]=λi=exiTβ

The log likelihood function is

lnL=∑i=1n[-λi+yixiTβ-lnyi!] Truncated Poisson distribution

Let y a discrete random variable follow Poisson distribution with mean and variance (λ) then the probability mass function of Poisson distribution is [19, 20]:

P(yi;λ)=e-λλyiyi!,   yi=0,1,2,…   , λ>0

A sample is considered truncated if the observations are limited to a certain portion of the population distribution. The truncated distribution is a subset of the untruncated distribution, with truncation occurring either from the left side [left truncated (y > l)] at a point, l, from the right side [right truncated (y < k)] at a point, k, or from both sides within the interval [l, k]. The probability density function of the shortened random variable can be described as a condition a distribution, as demonstrated below [19, 21]:

Case (1): left truncated Poisson at zero (yi > 0)

The zero-truncated Poisson distribution (ZTP) was first introduced by David and Johnson [36], which is distributed as y ~ ZTP(λ) and it is one of the models of logarithmic linear regression. The probability mass function of the zero truncated Poisson (ZTP) distribution is [3, 22, 23]:

P(yi|yi >0)=P(yi;λ)Pr[yi>0]                            = e-λλyiyi!1-Pr[y=0]                            = e-λλyi(1-e-λ)yi!, yi=1,2,3,…

Case (2): right truncated Poisson (yi ≤ k)

The Poisson distribution becomes right-truncated when truncation occurs at (k) where (yi ≤ k). The probability mass function of the right truncated Poisson distribution takes the following form [24, 25]:

P(yi|yi≤k )=P(yi;λ)Pr[yi≤k]                            =e-λλyiyi! (∑z=0ke-λλzz!)                            =λyi(∑z=0kλzz!)yi! , yi=0,1,2,…,k

Case (3): double truncated Poisson (l ≤ yi ≤ k)

The Double truncated Poisson data result from merging left truncated and right truncated Poisson data types. The probability mass function of the double truncated Poisson distribution takes the following form [26].

P(yi|l≤yi≤k)=P(xi;λ)Pr[l≤yi≤k]                            =λyi(∑z=lkλzz!)yi! , yi=l,l+1,l+2,…,k

The analysis of right truncation and double truncation for count data has received less scholarly focus compared to left truncation. One possible explanation is that Left-truncation occurs more frequently than right-truncation [26].

 Truncated Poisson regression model

The truncated Poisson regression model is one of the models of logarithmic linear regression for the dependent variable (yi) and is defined by the following formula [19, 21, 27, 28]:

yi=exiTβ+Ui   U~P(λ)

where  xiTβ=β0+β1xi1+…+ βnxin

yi~P(λ)

The distribution parameter of the response variable (yi) can be expressed as [29–31]:

λi=exiTβlnλi=xiTβ

In The truncated Poisson regression model the observations of (yi, xi) are obtained only for part of the population. The main goal of regression analysis involves parameter estimation to understand the relationship between dependent variable and independent variables. The maximum likelihood estimator serves to calculate parameter estimates for truncated Poisson regression models.

 Maximum likelihood estimation (MLE)

The estimation of parameters represents a fundamental research topic that attracts mathematical statistics researchers, because new estimation methods require accurate parameter estimation and optimal estimator identification [32].

Case (1): zero truncated Poisson regression model

The maximum likelihood function for the zero truncated Poisson regression model derives from the conditional probability function of the zero truncated Poisson distribution shown in Equation 6.

             L(β|y)=Πi=1ne-λλyi(1-e-λ)yi!ln L(β|y)=∑i=1n[yiln(λ)-λ-ln(1-e-λ)-ln(yi!)]ln L(β|y)=∑i=1n[yiβ-exiTβ-ln(1-e-xiTβ)-ln(yi!)]

The zero truncated Poisson maximum likelihood estimators require the derivative of Equation 10 with respect to β to obtain their values as:

∂lnL∂β=∑i=1n[yi-exiTβ(1-e-exiTβ)]xi=0

Case (2): right truncated Poisson regression model

The maximum likelihood function for the right truncated Poisson regression model derives from the conditional probability function of the right truncated Poisson distribution shown in Equation 7.

              L(β|y)=Πi=1nλyi(∑z=0kλzz!)yi!ln L(β|y)=∑i=1n[yiln(λ)-ln(yi!)-ln(∑z=0kλzz!)]ln L(β|y)=∑i=1n[yiβ-ln(yi!)-ln(∑z=0k(xiTβ)zz!)]

We derive maximum likelihood right truncated Poisson regression estimators by taking the first derivative of β and setting it to zero according to the following:

∂lnL∂β=∑i=1n[yi−∑z=0kxiz(xiTβ)zz!∑z=0k(xiTβ)zz!]=0

Case (3): double truncated Poisson regression model

The maximum likelihood function for the double truncated Poisson regression model derives from the Equation 8.

      L(β|y)=Πi=1nλyi(∑z=lkλzz!)yi!ln L(β|y)=∑i=1n[yiln(λ)-ln(yi!)-ln(∑z=lkλzz!)]ln L(β|y)=∑i=1n[yiβ-ln(yi!)-ln(∑z=lk(xiTβ)zz!)]

The maximum likelihood double truncated Poisson regression estimators can be determined differentiating Equation 14 with respect to β, giving

∂lnL∂β=∑i=1n[yi−∑z=lkxiz(xiTβ)zz!∑z=lk(xiTβ)zz!]=0

Equations 11, 13, 15 contain non-linear relationships between parameters which require iterative methods including Newton Raphson or Fisher scoring. In this research, we employ the traditional optimization method [Quasi newton method (BFGS)] alongside meta heuristic algorithms to estimate parameters of truncated Poisson regression model in three different cases, as well as improving a new algorithm to enhance the solution process.

 Broyden–Fletcher–Goldfarb– Shanno Method (BFGS)

The BFGS method is one of the quasi-Newton algorithms for unconstrained optimization problem, for finding a point x* ∈ Rn let:

minf(x)  x*∈Rn

Where the objective function f:Rn → R is a twice continuously differentiable objective function. The Broyden–Fletcher–Goldfarb–Shanno Method (BFGS) performs an iterative process that follows this procedure [13, 33]:

Start with x1, an initial point and [H1], a positive definite symmetric matrix n × n. where [H1] is the identity matrix [I]. Set i = 1 the iteration number.

Determine the gradient ∇fi at point x1, then set:

Si=-[H1]∇fi

Determine the optimal step length λi* moving along direction Si, then set:

xi+1=xi+λi*Si

Check if the new point xi+1 represents an optimal solution. If xi+1 is optimal, stop. Otherwise, go to step 5.

Update the matrix [H1] as:

[Hi+1]=[Hi]+(1+giT[Hi]gidiTgi)×didiTdiTgi-digiT[Hi]diTgi-[Hi]gidiTdiTgi

Where

gi=∇f(xi+1)- ∇f(xi)= ∇fi+1-∇fi

di=xi+1- xi

Set i = i + 1 the new iteration number, and go to step 2.

 Black winged kites algorithm (BKA)

In (2024) Wang et al. [15] introduced the black-winged kite optimization algorithm (BKA) which represents a groundbreaking meta-heuristic algorithm. The black-winged kite uses its survival techniques as the basis for its optimization algorithm. This bird utilizes excellent hovering capabilities in addition to its unexpected hunting proficiency. The black-winged kite feeds on insects together with birds and reptiles and small mammals. A model was developed through analysis of black-winged kite movement patterns and hunting abilities [15, 34, 35].

Initialization

The first step in BKA requires generating random solutions to establish the population as depicted in Algorithm 1. Each Black-winged kite receives its position through a uniform distribution:

Xi=BKlb+rand(BKub-BKlb)

Where BKlb, BKub represent the lower and upper bounds of ith black kites, respectively, and rand ∈ [0, 1] is a random number [15, 34, 35].

Black winged kite algorithm.

Attacking

The black-winged kite waits quietly before dropping to attack its prey after matching its wings and tail to the wind speed. The black-winged kite underwent two different assault scenarios throughout the global exploration phase of the BKA. The kite maintains its hovering position in the air while it readjusts its position to reach the target at its optimal attack angle. The kite maintains its position in the air while scanning for targets before striking down the most vulnerable one it detects. The attack behavior model uses the following mathematical expression:

xt+1i,j={xti,j+n(1+sin(r))×xti,jp<rxti,j+n×(2r-1)×xti,jelse n=0.05×e-2*(tT)2

Where xt+1i,j, xti,j is the position of the ith Black-winged kites in the jth dimension at iteration steps (t) and (t + 1)th, respectively. r ∈ [0, 1] is a random number and p = 0.9 is a constant value. T represents the total number of iterations and t is the current iteration [15, 34, 35].

 Migration

Bird migration occurs as an intricate behavior because both climate conditions and food availability serve as influential environmental elements. Bird migration exists as an adaptation to seasonal changes through which numerous birds move from northern regions to southern areas for improved living conditions and resources. Migration teams follow leaders who need excellent navigation abilities to achieve success. Our hypothesis relies on bird migration principles which state that if the fitness value of the current population is lower than that of the random population, the leader will give up leadership and join the migratory population, indicating that it is not suitable to lead the population forward. On the other hand, if the fitness value of the current population is higher than that of the random population, then the population will be guided to its destination. The approach enables automatic selection of superior leaders to achieve migration success. The mathematical model describes for the migration patterns of black-winged kites as follows:

xt+1i,j={xti,j+C(0,1)×(xti,j-Ltj)Fi<Frixti,j+C(0,1)×(Ltj-m×xti,j)else m=2×sin(r+π/2)

The parameter (m) is used to scale the current position of the kite in the update term, and scales the step size toward the leader perturbed by Cauchy mutation.

Where Ltj is the leading scorer of the black-winged kites in the jth dimension of the tth iteration so far. Fi is the fitness value of the current position obtained by any black-winged kite in the jth dimension of the tth iteration. Fri is the fitness value of the random position obtained from any black kites in the jth dimension of the tth iteration, and C(0, 1) is the Cauchy mutation. The probability density function of the Cauchy distribution is [15, 34, 35]:

f(x,δ,μ)=1πδδ2+(x-μ)2 -∞<x<∞

When δ = 0, μ = 1, then the standard form of the Cauchy distribution becomes the following:

f(x,δ,μ)=1π1x2+1 -∞<x<∞ Proposed algorithm

In this algorithm, the classical optimization algorithm, BFGS, was combined with the black-winged kite optimization algorithm (BFGS-BKA). where the randomness and speed of the black-winged kite optimization algorithm are used to find the optimal step length λi* in each iteration, while using parameter values before truncated as initial values for BFGS algorithm. The basic steps of this algorithm can be describes as:

Step (1): start with x1, an initial point and [H1], a positive definite symmetric matrix n × n. where [H1] is the identity matrix [I]. Set i = 1 the iteration number.

Step (2): determine the gradient ∇fi at point x1, then set:

Si = −[H1]∇fi

Step (3): find the optimal step length λi* by black-winged kite optimization algorithm:

Randomly generate the initial population of black-winged kite between λlb and λub as λ1, λ2, ........, λN. Evaluate the fitness value of each black-winged kite as (λ1), f(λ2), ......, f (λN). Set t = 1 the iteration number.

BKA chooses the individual with the best fitness value to become the leader λL in the initial population.

Find fbest and λL, the black-winged kite algorithm initiates its global exploration and search during its attack behavior according the following equation:

λt+1i,j={λti,j+n(1+sin(r))×λti,jp<rλti,j+n×(2r-1)×λti,jelse

In Bird migration, will be the position is update based on the fitness value of the leader as:

λt+1i,j={λti,j+C(0,1)×(λti,j-Ltj)Fi<Friλti,j+C(0,1)×(Ltj-m×λti,j)else

Test the convergence of the current solution. If the convergence criterion is not satisfied, go to step (3) and update the iteration number as t = t + 1, until convergence occurs and the optimal value of λ is determined.

Step (4): find new point:

xi+1=xi+λi*Si

Step (5): check if the new point xi+1 represents an optimal solution. If xi+1 is optimal, stop. Otherwise, go to step 6.

Step (6): update the matrix [H1] as:

[Hi+1]=[Hi]+(1+giT[Hi]gidiTgi)didiTdiTgi-digiT[Hi]diTgi-[Hi]gidiTdiTgi

Where

gi=∇f(xi+1)- ∇f(xi)= ∇fi+1-∇fidi=xi+1- xi

Step (7): set i = i + 1 the new iteration number, and go to step 2.

 Real data

This section contains real dataset to demonstrate the empirical importance of the algorithm (BFGS-BKA), where the real dataset is for the zero truncated Poisson regression model study. We use the (Arizona MedPar database, 1991). The dataset contains 1,495 observations the response variable, Lose, represents length of hospital stay. The explanatory variables for this model include an indicator of White (Patient identifies themselves as Caucasian, binary), Hmo (Patient belongs to a Health Maintenance Organization, binary), Type2 (Urgent admission, binary) and Type3 (Elective admission, binary).

As illustrated by Table 1, across all four methods, the parameter estimates are remarkably consistent, indicating stable estimation despite the choice of optimization algorithm. The numerical differences between the methods on each parameter are very small showing that all methods converge to nearly identical solutions. The MSE measures the average squared difference between observed and predicted values, serving as an indicator of model fit quality and estimator accuracy. Newton Raphson, Davidon–Fletcher–Powell (DFP) method, and BFGS yield the same MSE value of 0.0225, whereas the BFGS-BKA method achieves a slightly lower MSE of 0.0211. This implies that integrating the BKA algorithm with BFGS potentially enhances estimation accuracy or predictive performance. This improvement indicates that BFGS-BKA may better navigate the complex truncated likelihood surface to find a more optimal parameter set or avoid local optima.

Results zero-truncated Poisson regression model for the medpar data.

Parameters Methods
Newton Raphson DFP BFGS BFGS-BKA
Constant 2.33286 2.33286 2.33286 2.33281
White −0.1539437 −0.1539437 −0.1539437 −0.1539429
Hmo −0.0716485 −0.0716485 −0.0716485 −0.0716484
Type2 0.2217806 0.2217807 0.2217806 0.2217802
Type3 0.7096162 0.7096164 0.7096162 0.7096157
MSE 0.0225 0.0225 0.0225 0.0211
Simulation study

This section presented a simulation study to demonstrate the empirical importance of the algorithm (BFGS-BKA) to evaluate the accuracy of MLEs of parameters estimation of a truncated Poisson regression model of three cases left truncated Poisson regression model (LTPRM), right truncated Poisson regression model (RTPRM) and double truncated Poisson regression model (DTPRM). We consider dimension P = 3, 7, 12 and the sample sizes n = 25, 50, 100, 200.

Simulation results of the left truncated Poisson regression model

This section represented results for fit the left truncated Poisson regression model, with responses truncated at L = 0. Where independent variables are randomly generated according to a normal distribution and the data were simulated with true values beta = [0.5; −0.3; 0.2], beta = [0.5; −0.3; 0.2; 0.1; −0.1; 0.3; −0.2] and beta = [0.5; −0.3; 0.2; 0.1; −0.1; 0.3; −0.2; 0.4; −0.25; 0.15; 0.05; −0.05]. Tables 2–4 show the simulation results for LTPRM.

Simulation results for estimating parameters of a left truncated Poisson regression model at zero (y > 0; P = 3).

Sample size (n) Parameters Methods
Newton Raphson DFP BFGS BFGS-BKA
25 β0 0.0605861 0.0606416 0.0606268 0.0606055
β1 0.9904165 0.9904273 0.9904202 0.9904173
β2 −0.524277 −0.5242166 −0.5242486 −0.5242437
β3 0.009349 0.0093336 0.0093399 0.0093384
MSE 0.1822 0.1804 0.1801 0.1786
50 β0 −0.2315287 −0.2315281 −0.2315132 −0.2315116
β1 0.7735502 0.7735496 0.7735485 0.7735473
β2 −0.3058557 −0.3058522 −0.3058719 −0.3058719
β3 0.3230417 0.3230426 0.3230253 0.3230234
MSE 0.1477 0.1459 0.1453 0.1418
100 β0 0.0433403 0.0433412 0.0433433 0.0433433
β1 0.5843068 0.5843067 0.5843021 0.5843008
β2 −0.4143445 −0.4143451 −0.4143427 −0.4143419
β3 0.0748848 0.0748851 0.0748824 0.0748806
MSE 0.0952 0.0947 0.0938 0.0915
200 β0 −0.0092981 −0.009296 −0.0092982 −0.0092982
β1 0.7450347 0.7450345 0.7450346 0.7450338
β2 −0.4905174 −0.4905167 −0.4905176 −0.4905176
β3 0.2223349 0.2223347 0.222335 0.2223267
MSE 0.0773 0.0769 0.0767 0.0745

The best results are in bold.

Simulation results for estimating parameters of a left truncated Poisson regression model at zero (y > 0; P = 7).

Sample size (n) Parameters Methods
Newton Raphson DFP BFGS BFGS-BKA
25 β0 0.0926978 0.0926949 0.0927003 0.0927015
β1 0.564459 0.564461 0.5644587 0.5644587
β2 −0.6440744 −0.6440738 −0.6440724 −0.6440724
β3 0.280506 0.2805056 0.2805057 0.2805057
β4 0.069465 0.069467 0.0694647 0.0694631
β5 0.1659867 0.1659801 0.1659896 0.1659896
β6 0.22054 0.2205372 0.2205399 0.2205399
β7 0.1552263 0.155215 0.1552279 0.1552279
MSE 0.2515 0.2509 0.2507 0.2484
50 β0 −0.3173069 −0.3173081 −0.3173909 −0.3173909
β1 0.655645 0.6556289 0.6557447 0.6557432
β2 −0.2523863 −0.252384 −0.2523919 −0.25239
β3 0.3195793 0.3195883 0.3195612 0.3195612
β4 0.1808037 0.1808076 0.1808408 0.1808403
β5 −0.0684644 −0.0684646 −0.0685744 −0.0685744
β6 0.3032056 0.303196 0.303276 0.3032741
β7 −0.2861774 −0.286179 −0.2862013 −0.2862007
MSE 0.2219 0.2216 0.2215 0.2201
100 β0 −0.1071277 −0.1071258 −0.1071276 −0.107126
β1 0.7374492 0.7374496 0.7374505 0.7374505
β2 −0.1614437 −0.161447 −0.1614488 −0.1614488
β3 0.3848424 0.3848399 0.3848526 0.3848512
β4 0.01762 0.0176189 0.0176141 0.0176123
β5 0.0574704 0.057468 0.0574725 0.0574725
β6 0.3110027 0.3110044 0.3110038 0.3110038
β7 −0.4387681 −0.438765 −0.4387697 −0.4387697
MSE 0.1042 0.1040 0.1039 0.1021
200 β0 −0.2050698 −0.2050574 −0.2050697 −0.2050696
β1 0.5836933 0.5836848 0.5836637 0.5836614
β2 −0.2069601 −0.2069578 −0.2069618 −0.2069618
β3 0.3581892 0.3581868 0.3581863 0.3581846
β4 0.0030445 0.0030671 0.0030308 0.0030302
β5 −0.202504 −0.2024983 −0.2025151 −0.2025151
β6 0.4400861 0.440103 0.4400701 0.4400701
β7 −0.2052846 −0.2052629 −0.2052789 −0.2052763
MSE 0.0945 0.0943 0.0941 0.0915

The best results are in bold.

Simulation results for estimating parameters of a left truncated Poisson regression model at zero (y > 0; P = 12).

Sample size (n) Parameters Methods
Newton Raphson DFP BFGS BFGS-BKA
25 β0 −0.8614353 −0.8614325 −0.861444 −0.861427
β1 0.7035994 0.7035977 0.7036011 0.7036014
β2 −0.1529654 −0.1529619 −0.1529639 −0.1529633
β3 0.48303 0.4830395 0.4830308 0.4830301
β4 −0.7203112 −0.7203052 −0.7203215 −0.7203212
β5 0.3923419 0.3923434 0.3923467 0.3923471
β6 0.0631014 0.0631054 0.0630985 0.0630964
β7 −0.0710695 −0.0710814 −0.0710717 −0.071072
β8 −0.8822538 −0.8822546 −0.8822576 −0.8822573
β9 0.0181201 0.0181137 0.0181281 0.018128
β10 −0.6478624 −0.6478539 −0.6478666 −0.6478666
β11 0.4385116 0.4385079 0.4385171 0.4385169
β12 −0.3494832 −0.3494796 −0.3494841 −0.3494845
MSE 0.3152 0.3149 0.3147 0.3124
50 β0 0.2450007 0.2450028 0.2450011 0.2450013
β1 0.5359116 0.5359177 0.5359103 0.5359106
β2 −0.4534861 −0.4534876 −0.4534873 −0.4534871
β3 0.1330119 0.1330141 0.1330123 0.1330123
β4 0.1638947 0.1638883 0.1638955 0.1638962
β5 −0.0805194 −0.0805205 −0.0805195 −0.0805197
β6 0.4945398 0.494542 0.4945397 0.4945393
β7 −0.1883591 −0.1883612 −0.1883582 −0.1883579
β8 0.4052023 0.4052011 0.4052019 0.4052012
β9 −0.059146 −0.0591436 −0.0591467 −0.0591471
β10 −0.0556521 −0.0556525 −0.0556508 −0.0556504
β11 −0.12719 −0.127192 −0.1271926 −0.1271935
β12 0.0566836 0.0566845 0.056682 0.0566878
MSE 0.2147 0.2144 0.2139 0.2112
100 β0 −0.118253 −0.1182646 −0.1182716 −0.1182834
β1 0.6130675 0.6130611 0.6130748 0.6130739
β2 −0.3620159 −0.3620285 −0.3620218 −0.3620213
β3 0.0938253 0.0938262 0.0938282 0.0938291
β4 0.1551607 0.1551616 0.1551665 0.1551687
β5 −0.1028906 −0.102884 −0.1028899 −0.1028899
β6 0.3558426 0.3558493 0.3558494 0.3558498
β7 −0.0969868 −0.0969838 −0.0969893 −0.0969895
β8 0.3841613 0.3841709 0.3841673 0.3841672
β9 −0.1864775 −0.1864835 −0.1864828 −0.1864822
β10 0.2535725 0.253565 0.2535692 0.2535697
β11 −0.2586034 −0.2586041 −0.2586022 −0.2586019
β12 −0.0180681 −0.0180704 −0.0180699 −0.0180699
MSE 0.1132 0.1128 0.1126 0.1113
200 β0 −0.2619105 −0.2619025 −0.2618937 −0.2618923
β1 0.5147301 0.5147283 0.5147267 0.5147252
β2 −0.3444794 −0.3444768 −0.3444744 −0.3444749
β3 0.1516148 0.1516168 0.1516111 0.1516112
β4 0.0727301 0.0727304 0.0727338 0.0727341
β5 −0.0577983 −0.0577959 −0.0577988 −0.0577988
β6 0.4405197 0.4405187 0.4405171 0.4405179
β7 −0.335039 −0.3350374 −0.3350321 −0.3350314
β8 0.4311549 0.4311539 0.4311521 0.4311528
β9 −0.1879824 −0.1879845 −0.1879841 −0.187984
β10 0.1386025 0.1386018 0.1386025 0.1386025
β11 0.0912978 0.0912964 0.0912963 0.0912961
β12 −0.05895 −0.0589494 −0.0589534 −0.0589531
MSE 0.0823 0.0821 0.0818 0.0802

The best results are in bold.

 Simulation results of the right truncated Poisson regression model

This section represented results for fit the right truncated Poisson regression model, with responses truncated at U = 5. Where independent variables are randomly generated according to a normal distribution using matlab randn (n,p), also the data were simulated with true values beta = [0.5; −0.3; 0.2], beta = [0.5; −0.3; 0.2; 0.1; −0.1; 0.3; −0.2] and beta = [0.5; −0.3; 0.2; 0.1; −0.1; 0.3; −0.2; 0.4; −0.25; 0.15; 0.05; −0.05]. Tables 5–7 show the simulation results for (RTPRM).

Simulation results for estimating parameters of a right truncated Poisson regression model (P = 3).

Sample size (n) Parameters Methods
Newton Raphson DFP BFGS BFGS-BKA
25 β0 −0.2044458 −0.2047354 −0.2045128 −0.2045012
β1 0.7133822 0.7136196 0.713473 0.7132561
β2 −0.6439646 −0.64431 −0.6440023 −0.6437815
β3 0.0267832 0.0265996 0.0266998 0.0266897
MSE 0.1829 0.1802 0.1801 0.1789
50 β0 0.1481886 0.1481877 0.1481885 0.1481875
β1 0.4981213 0.498122 0.4981215 0.4981203
β2 −0.5420273 −0.5420278 −0.5420273 −0.5420264
β3 0.3070644 0.3070648 0.3070653 0.3070621
MSE 0.1946 0.1931 0.1911 0.1909
100 β0 −0.0986433 −0.0986435 −0.0986435 −0.0986435
β1 0.5020687 0.5020689 0.5020689 0.5020689
β2 −0.338183 −0.338183 −0.3381831 −0.3381831
β3 0.1941197 0.1941197 0.1941197 0.1941197
MSE 0.0826 0.0813 0.0806 0.0801
200 β0 0.0338927 0.0338926 0.0338927 0.033893
β1 0.5449919 0.5449919 0.5449919 0.5449919
β2 −0.2905865 −0.2905865 −0.2905865 −0.2905865
β3 0.2194149 0.2194149 0.2194149 0.2194149
MSE 0.0716 0.0715 0.0714 0.0701

The best results are in bold.

Simulation results for estimating parameters of a right truncated Poisson regression model (P = 7).

Sample size (n) Parameters Methods
Newton Raphson DFP BFGS BFGS-BKA
25 β0 −0.4794763 −0.4794751 −0.4794751 −0.4794751
β1 0.5107009 0.5107017 0.5107012 0.5107006
β2 0.1826785 0.1826817 0.1826795 0.1826711
β3 0.3178249 0.3178251 0.3178251 0.3178251
β4 0.4424957 0.4424962 0.4424953 0.442494
β5 −0.3870675 −0.3870688 −0.3870697 −0.3870697
β6 0.8706261 0.8706264 0.8706269 0.8706269
β7 0.1193511 0.1193497 0.1193495 0.119349
MSE 0.2547 0.2545 0.2543 0.2532
50 β0 −0.047451 −0.0474496 −0.0474511 −0.0474511
β1 0.0796038 0.0796035 0.079604 0.079604
β2 −0.2130617 −0.2130617 −0.2130618 −0.2130618
β3 −0.1061506 −0.1061513 −0.1061506 −0.1061502
β4 0.6912239 0.6912238 0.691224 0.6912242
β5 −0.2849013 −0.2849013 −0.2849013 −0.2849013
β6 0.0200921 0.0200916 0.0200923 0.0200925
β7 −0.0268927 −0.0268926 −0.0268927 −0.0268927
MSE 0.2031 0.2027 0.2025 0.2012
100 β0 −0.0201553 −0.0201562 −0.0201553 −0.0201458
β1 0.3946759 0.3946762 0.3946759 0.3946742
β2 −0.1337339 −0.1337337 −0.1337339 −0.1337339
β3 0.0826245 0.0826248 0.0826245 0.0826245
β4 0.4280978 0.4280989 0.4280978 0.4280978
β5 −0.1179131 −0.1179137 −0.1179131 −0.1179131
β6 0.3534397 0.3534403 0.3534397 0.3534397
β7 −0.0139108 −0.0139105 −0.0139108 −0.0139108
MSE 0.1132 0.1131 0.1129 0.1118
200 β0 0.003164 0.003164 0.003164 0.003164
β1 0.3648677 0.3648678 0.3648677 0.3648669
β2 −0.1732149 −0.1732149 −0.1732149 −0.1732149
β3 0.0008671 0.0008671 0.0008671 0.0008671
β4 0.5622831 0.5622832 0.5622832 0.5622825
β5 −0.0834699 −0.0834699 −0.0834698 −0.0834696
β6 0.3784646 0.3784646 0.3784646 0.3784646
β7 0.1480245 0.1480245 0.1480245 0.1480245
MSE 0.0831 0.0827 0.0824 0.0813

The best results are in bold.

Simulation results for estimating parameters of a right truncated Poisson regression model (P = 12).

Sample size (n) Parameters Methods
Newton Raphson DFP BFGS BFGS-BKA
25 β0 0.5074047 0.5074047 0.5074047 0.5074047
β1 0.2848892 0.2848892 0.2848892 0.2848892
β2 −0.4693933 −0.4693933 −0.4693933 −0.4693933
β3 0.5707239 0.5707239 0.5707239 0.5707239
β4 0.0818152 0.0818152 0.0818152 0.0818152
β5 −0.1785226 −0.1785226 −0.1785226 −0.1785223
β6 0.9287227 0.9287227 0.9287227 0.9287227
β7 −0.4038411 −0.4038411 −0.4038411 −0.4038411
β8 0.3582214 0.3582214 0.3582214 0.3582214
β9 −0.1126234 −0.1126234 −0.1126234 −0.1126227
β10 0.5869459 0.5869459 0.5869459 0.5869459
β11 0.0692677 0.0692677 0.0692677 0.0692677
β12 −0.0358698 −0.0358698 −0.0358698 −0.0358698
MSE 0.3163 0.3161 0.3160 0.3141
50 β0 −0.4827961 −0.4828029 −0.4828315 −0.4828273
β1 0.4601612 0.460162 0.4601896 0.4601899
β2 −0.6072522 −0.6072504 −0.6072636 −0.6072651
β3 0.069667 0.0696696 0.0696657 0.0696643
β4 0.1243769 0.1243796 0.1244038 0.1244038
β5 −0.4199463 −0.419951 −0.4200035 −0.4200035
β6 0.3623213 0.3623205 0.3623135 0.3623116
β7 −0.2892603 −0.2892586 −0.2892488 −0.2892465
β8 0.9695849 0.9695876 0.9696129 0.9696129
β9 −0.0936884 −0.0936863 −0.0936845 −0.0936837
β10 0.612861 0.6128638 0.612854 0.612854
β11 0.1164395 0.1164401 0.1164395 0.1164395
β12 0.1428553 0.1428585 0.1428394 0.14283622
MSE 0.2338 0.2335 0.2332 0.2217
100 β0 −0.0069453 −0.0069457 −0.0069454 −0.0069454
β1 0.4531302 0.4531303 0.4531302 0.4531291
β2 −0.335599 −0.335599 −0.335599 −0.335599
β3 0.1139658 0.1139659 0.1139658 0.1139648
β4 0.0987151 0.0987151 0.0987151 0.0987151
β5 −0.1070247 −0.1070247 −0.1070247 −0.1070247
β6 0.1939749 0.1939748 0.1939749 0.1939749
β7 −0.3772053 −0.3772055 −0.3772054 −0.3772054
β8 0.491455 0.4914553 0.4914551 0.4914551
β9 −0.4061271 −0.4061273 −0.4061272 −0.406127
β10 0.0180528 0.018053 0.0180528 0.0180519
β11 0.0238667 0.0238668 0.0238668 0.0238668
β12 −0.2245255 −0.2245256 −0.2245255 −0.2245255
MSE 0.1255 0.1253 0.1251 0.1232
200 β0 −0.0483217 −0.0483214 −0.048323 −0.048323
β1 0.5104306 0.5104302 0.5104312 0.5104322
β2 −0.3227727 −0.3227724 −0.3227734 −0.3227734
β3 0.0754467 0.0754466 0.0754464 0.0754458
β4 0.0824297 0.0824298 0.08243 0.0824312
β5 −0.1066715 −0.1066716 −0.1066719 −0.106672
β6 0.2741968 0.2741964 0.2741966 0.2741965
β7 −0.2391153 −0.2391153 −0.2391158 −0.2391158
β8 0.49603 0.4960291 0.4960302 0.4960302
β9 −0.2210014 −0.2210012 −0.221002 −0.2210026
β10 0.2310687 0.2310681 0.2310689 0.2310689
β11 0.0333784 0.0333783 0.0333785 0.0333774
β12 −0.0125663 −0.0125661 −0.0125661 −0.0125661
MSE 0.0945 0.0942 0.0939 0.0921

The best results are in bold.

 Simulation results of the double truncated Poisson regression model

This section represented results for fit the double truncated Poisson regression model, with responses truncated at L = 0 and U = 5. Where Independent variables are randomly generated according to a normal distribution using matlab randn (n,p), also the data were simulated with true values beta = [0.5; −0.3; 0.2], beta = [0.5; −0.3; 0.2; 0.1; −0.1; 0.3; −0.2] and beta = [0.5; −0.3; 0.2; 0.1; −0.1; 0.3; −0.2; 0.4; −0.25; 0.15; 0.05; −0.05]. Tables 8–10 show the simulation results for (DTPRM).

Simulation results for estimating parameters of a double truncated Poisson regression model (P = 3).

Sample size (n) Parameters Methods
Newton Raphson DFP BFGS BFGS-BKA
25 β0 −0.1558273 −0.1557939 −0.1557566 −0.1557506
β1 0.4512201 0.4511945 0.4511494 0.4511398
β2 0.1766003 0.1766157 0.17659 0.1765241
β3 0.4351908 0.4351774 0.4351864 0.4351864
MSE 0.2955 0.2821 0.2763 0.2551
50 β0 −0.2589921 −0.2589981 −0.2590017 −0.2590017
β1 0.4438434 0.4438345 0.4438293 0.4438229
β2 0.4488448 0.4488376 0.4488351 0.4488305
β3 0.5731984 0.5731992 0.573202 0.573202
MSE 0.2331 0.2286 0.2147 0.2131
100 β0 −0.1156143 −0.115618 −0.1156356 −0.1156356
β1 0.3477845 0.3477863 0.3477778 0.3477746
β2 0.3157024 0.3157139 0.3157184 0.3157184
β3 0.2233149 0.2233078 0.2233294 0.2233267
MSE 0.1621 0.1618 0.1605 0.1587
200 β0 −0.0226799 −0.022708 −0.0226883 −0.0226848
β1 0.517915 0.5179045 0.5179474 0.5179462
β2 0.3185699 0.3185493 0.3185984 0.3185983
β3 0.4484079 0.4483947 0.4484859 0.4484837
MSE 0.1159 0.1077 0.1054 0.1018

The best results are in bold.

Simulation results for estimating parameters of a double truncated Poisson regression model (P = 7).

Sample size (n) Parameters Methods
Newton Raphson DFP BFGS BFGS-BKA
25 β0 −0.2858235 −0.2857848 −0.2858225 −0.2858223
β1 −0.1653242 −0.1653189 −0.1653308 −0.1653312
β2 0.2407589 0.2407697 0.240767 0.2407665
β3 0.8443157 0.8442951 0.8442992 0.8442997
β4 0.3714291 0.3714186 0.3714336 0.3714342
β5 −0.2360293 −0.2360385 −0.236036 −0.2360356
β6 0.3906935 0.3906461 0.3906882 0.3906893
β7 −0.5665927 −0.566663 −0.5665713 −0.5665721
MSE 0.3215 0.3208 0.3195 0.3174
50 β0 −0.0375944 −0.0375962 −0.0376234 −0.0376282
β1 0.66365 0.6636705 0.6636389 0.663639
β2 −0.1675661 −0.1675437 −0.1675875 −0.1675894
β3 0.345409 0.3453086 0.3454378 0.3454386
β4 −0.0118677 −0.0118937 −0.0117899 −0.011789
β5 0.0154284 0.0154216 0.0154415 0.0154432
β6 0.258271 0.2583602 0.2583117 0.2583148
β7 −0.3189341 −0.3189986 −0.3189528 −0.3189539
MSE 0.3081 0.3025 0.3018 0.3004
100 β0 −0.2282489 −0.2282489 −0.228258 −0.228242
β1 0.3241313 0.3241313 0.3241238 0.3241238
β2 −0.1384214 −0.1384214 −0.13842 −0.138367
β3 0.5801789 0.5801789 0.5801797 0.5801854
β4 0.0858746 0.0858746 0.0858833 0.0858864
β5 −0.0348083 −0.0348083 −0.0348339 −0.0348425
β6 0.5331769 0.5331769 0.533192 0.5331974
β7 −0.2537133 −0.2537133 −0.2537146 −0.2537185
MSE 0.2137 0.2136 0.2041 0.2028
200 β0 −0.1587123 −0.1587184 −0.1587097 −0.1587068
β1 0.6670115 0.6670268 0.667012 0.6670125
β2 −0.5487419 −0.5487332 −0.5487403 −0.5487401
β3 0.3911965 0.3911697 0.3911951 0.3911955
β4 −0.0013219 −0.0013251 −0.0013203 −0.00132
β5 −0.0321376 −0.0321579 −0.0321368 −0.0321347
β6 0.4452084 0.4451819 0.4452131 0.4452159
β7 −0.1397936 −0.1398144 −0.1397963 −0.1397963
MSE 0.1486 0.1337 0.1314 0.1289

The best results are in bold.

Simulation results for estimating parameters of a double truncated Poisson regression model (P = 12).

Sample size (n) Parameters Methods
Newton Raphson DFP BFGS BFGS-BKA
25 β0 −0.6073089 −0.6073912 −0.607189 −0.607152
β1 −0.1323503 −0.1323735 −0.1323581 −0.1323575
β2 0.3691619 0.3691301 0.3690593 0.3690581
β3 −0.259072 −0.2590905 −0.259066 −0.2590571
β4 −0.1889251 −0.1889383 −0.1888898 −0.1888643
β5 −0.5384405 −0.5384847 −0.538497 −0.538499
β6 0.6892821 0.6893039 0.6893448 0.6893673
β7 −0.3449217 −0.3449529 −0.344885 −0.344839
β8 −0.2177979 −0.217815 −0.2178416 −0.2178625
β9 −0.5734173 −0.5734628 −0.5733326 −0.5733308
β10 −0.3854379 −0.3854759 −0.385406 −0.385392
β11 1.314073 1.314133 1.313973 1.313947
β12 1.279418 1.279486 1.279205 1.279201
MSE 0.4132 0.4085 0.4055 0.4031
50 β0 0.0260518 0.0259643 0.0260443 0.0260427
β1 0.2475601 0.2475584 0.2475476 0.2475293
β2 −0.2577356 −0.2577761 −0.2577354 −0.2577316
β3 0.0468957 0.0469244 0.0468872 0.0468621
β4 −0.1204874 −0.1204298 −0.1204943 −0.1204997
β5 −0.4188021 −0.4188395 −0.4187998 −0.4187205
β6 0.047606 0.0475921 0.0476052 0.0476044
β7 −0.0003571 −0.0002965 −0.0003472 −0.0003472
β8 0.1689888 0.1689627 0.1689886 0.1689881
β9 −0.9068353 −0.9069006 −0.9068387 −0.9068392
β10 0.178105 0.178127 0.1781101 0.1781113
β11 0.4888108 0.4888359 0.4888204 0.4888252
β12 0.3629748 0.3630487 0.3629835 0.3629927
MSE 0.3462 0.3255 0.3032 0.3001
100 β0 0.0385659 0.0385516 0.0385577 0.0385582
β1 0.4102414 0.410243 0.4102418 0.4102437
β2 −0.6420943 −0.642092 −0.6421023 −0.6421035
β3 0.0022714 0.0022528 0.0022711 0.0022742
β4 −0.2230076 −0.2230055 −0.2230088 −0.2230089
β5 −0.1790879 −0.1790891 −0.1790935 −0.1790964
β6 0.355542 0.3555414 0.3555464 0.3555483
β7 −0.1333147 −0.133325 −0.1333074 −0.1333055
β8 0.4505356 0.4505381 0.4505453 0.4505461
β9 −0.4601778 −0.460187 −0.4601826 −0.4601815
β10 0.4124141 0.4124227 0.4124199 0.4124176
β11 0.1994184 0.1994095 0.1994132 0.1994264
β12 −0.1919182 −0.1919244 −0.1919197 −0.1919199
MSE 0.2046 0.2039 0.2008 0.1991
200 β0 −0.1014937 −0.1014946 −0.1015082 −0.1015153
β1 0.4214451 0.4214462 0.4214523 0.4214592
β2 −0.3636477 −0.3636489 −0.3636515 −0.3636594
β3 −0.2367109 −0.2367112 −0.2367138 −0.2367182
β4 −0.1965457 −0.1965465 −0.19655 −0.1965528
β5 −0.2376769 −0.2376772 −0.2376771 −0.2376771
β6 0.2626723 0.2626748 0.2626743 0.262674
β7 −0.1748725 −0.1748717 −0.1748697 −0.1748641
β8 0.5248844 0.5248844 0.5248921 0.5248938
β9 −0.3120583 −0.3120575 −0.3120569 −0.3120554
β10 0.1084512 0.108451 0.1084503 0.1084489
β11 0.0542112 0.0542105 0.0542091 0.0542063
β12 −0.0990921 −0.0990901 −0.0990931 −0.0990942
MSE 0.1387 0.1255 0.1208 0.1135

The best results are in bold.

From all the Tables above, we can concluded that across all sample sizes, parameter estimates from all four methods are highly similar, demonstrating consistent convergence to approximately the same values. This consistency suggests each optimization method is capable of locating reliable parameter estimates even in smaller samples. As sample size increases, the parameter estimates stabilize and vary less across methods. This reflects the expected property of maximum likelihood estimators: increased sample size yields more precise and stable estimates. As sample size increases (from 25 to 200), the estimated MSE decreases monotonically for all methods, demonstrating improved estimation precision with larger data consistent with statistical theory. BFGS-BKA consistently achieves the lowest MSE at each sample size, indicating superior estimation accuracy. Classical methods such as Newton-Raphson, DFP, and BFGS provide similar MSE values, slightly higher than BFGS-BKA. The performance gap, though sometimes small in absolute terms, illustrates the advantage of using the Black Kite Optimization (BKA) to enhance the classical BFGS method, improving optimization over the complex truncated likelihood surface. The similarity of parameter estimates across methods confirms robustness of the numerical algorithms when applied to truncated Poisson regression models.

Integrating the Black Kite Optimization algorithm with BFGS consistently improves the optimization process, yielding more accurate and stable estimates. This suggests that metaheuristic approaches like BKA help avoid local optima and improve convergence speed when dealing with truncated likelihood functions. For small sample sizes where numerical instability and local optima are more problematic, BFGS-BKA provides meaningful improvements. For larger datasets, while all methods perform well, BFGS-BKA still maintains a measurable edge in accuracy.

Figures 1–3 show the performance of Newton's, DFP, BFGS, and BFGS-BKA algorithms in terms of time that represents the average computation time (in seconds) used by the all algorithms in the case of successful runs. We can see how the proposed algorithm is better than the others because of the Cauchy mutation that causes better exploration and the leader strategy that causes faster convergence leading to a balance between global and local search.

Computational time of left truncated Poisson regression model at zero (y > 0) (a) P = 3, (b) P = 7 and (c) P = 12.

Three bar graphs compare the execution time, in seconds, of different methods for various sample sizes. Graph (a) compares the Newton, DFP, BFGS, and BFGS-BKA methods for P=3; graph (b) does the same for P=7; and graph (c) for P=12. In each graph, the methods are represented by different colored bars, showing trends and differences in execution time as sample size increases from 25 to 200.

Computational time of right truncated Poisson regression model (a) P = 3, (b) P = 7 and (c) P = 12.

Three bar charts compare execution times of four optimization methods (Newton, DFP, BFGS, BFGS-BKA) over different sample sizes. Chart (a) for P=3, (b) for P=7, and (c) for P=12, with execution time increasing consistently as sample sizes increase from 25 to 200. Each method's performance is visually distinguished by color.

Computational time of double truncated Poisson regression model (a) P = 3, (b) P = 7 and (c) P = 12.

Three bar charts compare execution times using four methods: Newton, DFP, BFGS, and BFGS-BKA. Chart (a) shows results for P=3; (b) for P=7; (c) for P=12. Execution time increases with sample size across all methods. Conclusions

This paper addressed the critical challenge of accurately estimating parameters in truncated Poisson regression models, where standard maximum likelihood estimation is complicated by a truncated likelihood function that includes a non-trivial normalizing constant. The paper proposed the use of the BKO algorithm, a metaheuristic inspired by the hunting and migratory behavior of black kites, which aims to enhance exploration and exploitation capabilities when searching the parameter space. By leveraging BKO, the study seeks to improve the stability, convergence speed, and accuracy of parameter estimates in truncated Poisson models. Simulation studies and empirical analyses included in the paper demonstrate the superior performance of this approach compared to classical optimization methods, highlighting its potential as a robust and efficient solution for parameter estimation in truncated count data settings.

 Data availability statement

The data analyzed in this study is subject to the following licenses/restrictions: the data will be under request. Requests to access these datasets should be directed to Zakariya Algamal, zakariya.algamal@uomosul.edu.iq.

 Author contributions

GB: Conceptualization, Formal analysis, Software, Writing – original draft, Writing – review & editing. SW: Conceptualization, Formal analysis, Methodology, Resources, Writing – original draft, Writing – review & editing. ZA: Supervision, Validation, Writing – review & editing.

 Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

 Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

 Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

 References Ahmed Alangood HN Algamal Z Khaleel MA. Variable Selection in Poisson Regression Model based on Chaotic Meta-Heuristic Search Algorithm. BIO Web of Conferences, 97 (2024). doi: 10.1051/bioconf/20249700161 Algamal Z. Variable Selection in Count Data Regression Model based on Firefly Algorithm. Statistics, Optimization and Information Computing, 7 (2019). doi: 10.19139/soic.v7i2.566 Alwani ZIZ Ibrahim AI Yunus RM Yusof F. Application of zero-truncated count data regression models to air-pollution disease. J Phys Conf Ser. (2021) 1988:012096. doi: 10.1088/1742-6596/1988/1/012096 Hilbe J-M. Modeling-Count-Data. Cambridge: Cambridge University Press (2014). doi: 10.1017/CBO9781139236065 Hawa NS Mustafa MY Kibria BMG Algamal ZY. Bootstrap Liu-type estimator for Conway-Maxwell-Poisson regression model. Commun Stat Simul Comput. (2025) 1–12. doi: 10.1080/03610918.2025.2462680 Kara R Yeşilova A. Zero truncated models in regression analysis: an examination of their advantages on small mean values. J Inst Nat Appl Sci. (2025) 30:102–12. doi: 10.53433/yyufbed.1590611 Alkhateeb A Algamal Z. Jackknifed liu-type estimator in Poisson regression model. J Iran Stat Soc. (2020) 19:21–37. doi: 10.29252/jirss.19.1.21 Liu Y Li W Zhang X. A marginalized zero-truncated Poisson regression model and its model averaging prediction. Commun Math Stat. (2025) 13:527–70. doi: 10.1007/s40304-022-00312-8 Wani MK Ahmad PB. One-inflated zero-truncated Poisson distribution: statistical properties and real life applications. Ann Data Sci. (2025) 12:639–66. doi: 10.1007/s40745-024-00526-3 Young DS Roemmele ES Yeh P. Zero-inflated modeling part I: traditional zero-inflated count regression models, their applications, and computational tools. WIREs Comput Stat. (2022) 14:e1541. doi: 10.1002/wics.1541 Qi Y Yu-Zhu T Yi-Jing Z Yue W Zhi-Bao M. An ordinal collaboration network model with zero truncated poisson latent variables and its application. Stat (2025) 14:e70040. doi: 10.1002/sta4.70040 Ünlü HK Young DS Yigiter A Hilal Özcebe L. A mixture model with Poisson and zero-truncated Poisson components to analyze road traffic accidents in Turkey. J Appl Stat. (2022) 49:1003–17. doi: 10.1080/02664763.2020.184361035707813 Rao SS. Engineering Optimization Theory and Practice. Hoboken, NJ: John Wiley & Sons, Inc. (2009). Yu X Gen M. Introduction to Evolutionary Algorithms. London: Springer-Verlag London Limited (2010). doi: 10.1109/ICCIE.2010.5668407 Wang J Wang WC Hu XX Qiu L Zang HF. Black-winged kite algorithm: a nature-inspired meta-heuristic for solving benchmark functions and engineering problems. Artif Intell Rev. (2024) 57:98. doi: 10.1007/s10462-024-10723-4 Zhou H Alexander D Lange K. A quasi-Newton acceleration for high-dimensional optimization algorithms. Stat Comput. (2011) 21:261–73. doi: 10.1007/s11222-009-9166-321359052 Jameel MS Basheer GT Al-Bayati AY Algamal ZY. Parameter estimation of a truncated regression model based on improving numerical optimization algorithms. J Phys Conf Ser. (2021) 1897:012059. doi: 10.1088/1742-6596/1897/1/012059 Hwang WH Stoklosa J Wang CY. Population size estimation using zero-truncated Poisson regression with measurement error. J Agric Biol Environ Stat. (2022) 27:303–20. doi: 10.1007/s13253-021-00481-z35813491 Green W. Econometric Analysis. 5th ed. Upper Saddle River, NJ: Prentice Hall (2003). Demaris A. Regression with Social Data Modeling Continuous and Limited Response Variables. New York, NY: John Wiley & Sons, Inc. (2004). doi: 10.1002/0471677566 Heij C de Boer P Franses PH Kloek T van Dijk HK. Econometric Methods with Applications in Business and Economics. Oxford: Oxford University Press (2004). Umar MA Jimoh K Yahya WB. A Note on the Applications of some Zero Truncated. Professional Statisticians Society of Nigeria, 3 (2019). Li X Sun Y Tian G Liang J Shi J. Mean regression model for the zero-truncated Poisson distribution and its generalization. Comput Stat Data Anal. (2023) 179:107650. doi: 10.1016/j.csda.2022.107650 Rahman S. Truncated Distributions and their Applications, 2004-2005 (2004). Mwandigha LM Fraser KJ Racine-Poon A Mouksassi MS Ghani AC. Power calculations for cluster randomized trials (CRTs) with right-truncated Poisson-distributed outcomes: a motivating example from a malaria vector control trial. Int J Epidemiol. (2020) 49:954–62. doi: 10.1093/ije/dyz27732011684 Suaiee AM. A Double Truncated Poisson Regression Model with Random Effects. Greeley, CO: University of Northern Colorado (2013). Karlsson M. Estimators of regression parameters for truncated and censored data. Metrika (2006) 63:329–41. doi: 10.1007/s00184-005-0023-x Newey WK. Conditional moment restrictions in censored and truncated regression models. Econ Theory (2001) 17:863–88. doi: 10.1017/S0266466601175018 Irshad M Chesneau C Shibu DS Monisha M Maya R. Lagrangian zero truncated Poisson distribution: properties regression model and applications. Symmetry (2022) 14:1775. doi: 10.3390/sym14091775 Sabri Al-zubaidi M Abdulah EK. Comparison of estimation methods for zero truncated poisson regression model. J Econ Admin Sci. (2024) 30:492–508. doi: 10.33095/7bt1f714 Kumar S Singh BP Madhusudan JV. Factors influencing migration: application of truncated poisson regression. Int J Adv Sci Technol. (2020) 29:9112–20. Hussain EA Al-Shallawi ANS Saied HA. Using maximum likelihood method to estimate parameters of the linear regression T truncated model. NTU J Pure Sci. (2022) 4:26–34. doi: 10.56286/ntujps.v1i4.313 Luksan L Spedicatob E. Variable metric methods for unconstrained optimization. J Comput Appl Math. (2000) 124:61–95. doi: 10.1016/S0377-0427(00)00420-9 Zhang Z Wang X Yue Y. Heuristic optimization algorithm of black-winged kite fused with osprey and its engineering application. Biomimetics (2024) 9:595. doi: 10.3390/biomimetics910059539451801 Zhao M Su Z Hua Z Zhao C. Improved black-winged kite algorithm based on chaotic mapping and adversarial learning. J Phys Conf Ser. (2024) 2898:012040. doi: 10.1088/1742-6596/2898/1/012040 David FN Johnson NL. The truncated poisson. Biometrics. (1952) 8:275–85.

Edited by: Appanah Rao Appadu, University of the Western Cape, South Africa

Reviewed by: Yannick Tangman, University of Mauritius, Mauritius

Faiza Sami, Govt Gordon Graduate College Rawalpindi, Pakistan