Atıf Ahmet Evren
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Atıf Ahmet Evren has a total of 11 co-authors. Here are the top co-authors by number of shared publications:
- Erhan Ustaoğlu
Department of Informatics, Faculty of Management, Marmara University, 34180 Istanbul, Turkey
4 shared publications - Elif Tuna
Department of Statistics, Faculty of Sciences and Literature, Yildiz Technical University, Davutpasa, Esenler, 34210 Istanbul, Turkey
3 shared publications - Zehra Zeynep Şahinbaşoğlu
Department of Statistics, Faculty of Sciences and Literature, Yildiz Technical University, Davutpasa, Esenler, 34210 Istanbul, Turkey
3 shared publications - Büşra Şahin
Department of Computer Engineering, Faculty of Engineering, Halic University, Eyupsultan, 34060 Istanbul, Turkey
2 shared publications - Mert Veznikli
Department of Biostatistics, School of Medicine, Koc University, 34450 Istanbul, Turkey
2 shared publications - Elif Tuna
Yıldız Teknik Üniversitesi
1 shared publication
Latest Publications
- Open Access
LAFS: A Fast, Differentiable Approach to Feature Selection Using Learnable Attention
Feature selection is a critical preprocessing step for mitigating the curse of dimensionality in machine learning. Existing methods present a difficult trade-off: filter methods are fast but often suboptimal as they evaluate features in isolation, while wrapper methods are powerful but computationally prohibitive due to their iterative nature. In this paper, we propose LAFS (Learnable Attention for Feature Selection), a novel, end-to-end differentiable framework that achieves the performance of wrapper methods at the speed of simpler models. LAFS employs a neural attention mechanism to learn a context-aware importance score for all features simultaneously in a single forward pass. To encourage the selection of a sparse and non-redundant feature subset, we introduce a novel hybrid loss function that combines the standard classification objective with an information-theoretic entropic regularizer on the attention weights. We validate our approach on real-world high-dimensional benchmark datasets. Our experiments demonstrate that LAFS successfully identifies complex feature interactions and handles multicollinearity. In general comparison, LAFS achieves very close and accurate results to state-of-the-art RFE-LGBM and embedded FSA methods. Our work establishes a new point on the accuracy-efficiency frontier, demonstrating that attention-based architectures provide a compatible solution to the feature selection problem.
Open AccessLAFS: A Fast, Differentiable Approach to Feature Selection Using Learnable Attention
Feature selection is a critical preprocessing step for mitigating the curse of dimensionality in machine learning. Existing methods present a difficult trade-off: filter methods are fast but often suboptimal as they evaluate features in isolation, while wrapper methods are powerful but computationally prohibitive due to their iterative nature. In this paper, we propose LAFS (Learnable Attention for Feature Selection), a novel, end-to-end differentiable framework that achieves the performance of wrapper methods at the speed of simpler models. LAFS employs a neural attention mechanism to learn a context-aware importance score for all features simultaneously in a single forward pass. To encourage the selection of a sparse and non-redundant feature subset, we introduce a novel hybrid loss function that combines the standard classification objective with an information-theoretic entropic regularizer on the attention weights. We validate our approach on real-world high-dimensional benchmark datasets. Our experiments demonstrate that LAFS successfully identifies complex feature interactions and handles multicollinearity. In general comparison, LAFS achieves very close and accurate results to state-of-the-art RFE-LGBM and embedded FSA methods. Our work establishes a new point on the accuracy-efficiency frontier, demonstrating that attention-based architectures provide a compatible solution to the feature selection problem.
- Open Access
Parameter Estimation of the Dirichlet Distribution Based on Entropy
The Dirichlet distribution as a multivariate generalization of the beta distribution is especially important for modeling categorical distributions. Hence, its applications vary within a wide range from modeling cell probabilities of contingency tables to modeling income inequalities. Thus, it is commonly used as the conjugate prior of the multinomial distribution in Bayesian statistics. In this study, the parameters of a bivariate Dirichlet distribution are estimated by entropy formalism. As an alternative to maximum likelihood and the method of moments, two methods based on the principle of maximum entropy are used, namely the ordinary entropy method and the parameter space expansion method. It is shown that in estimating the parameters of the bivariate Dirichlet distribution, the ordinary entropy method and the parameter space expansion method give the same results as the method of maximum likelihood. Thus, we emphasize that these two methods can be used alternatively in modeling bivariate and multinomial Dirichlet distributions.
Open AccessParameter Estimation of the Dirichlet Distribution Based on Entropy
The Dirichlet distribution as a multivariate generalization of the beta distribution is especially important for modeling categorical distributions. Hence, its applications vary within a wide range from modeling cell probabilities of contingency tables to modeling income inequalities. Thus, it is commonly used as the conjugate prior of the multinomial distribution in Bayesian statistics. In this study, the parameters of a bivariate Dirichlet distribution are estimated by entropy formalism. As an alternative to maximum likelihood and the method of moments, two methods based on the principle of maximum entropy are used, namely the ordinary entropy method and the parameter space expansion method. It is shown that in estimating the parameters of the bivariate Dirichlet distribution, the ordinary entropy method and the parameter space expansion method give the same results as the method of maximum likelihood. Thus, we emphasize that these two methods can be used alternatively in modeling bivariate and multinomial Dirichlet distributions.
- Open Access
- 25 May 2023
- Published by MDPI AG in Sustainability
- Vol. 15 (11) , 8569
Examination of the Effects of COVID-19 on Happiness in Different Geographical Regions with Piecewise Linear Panel Data Models
The COVID-19 pandemic has recently caused the loss of millions of lives, and billions of others have been deeply affected. This crisis has changed the way people live, think about life, and perceive happiness. The aim of this study is to reveal differences between geographical regions by investigating the effect of the happiness variable on different countries during the international COVID-19 pandemic. The primary purpose is to demonstrate how such a pandemic may affect different countries in terms of happiness at the individual level and to identify possible strategies for the future. With this aim, both static and dynamic panel data models were used while applying fixed effects, random effects, and the generalized method of moments (GMM). A basic assumption in panel data models is that the coefficients do not change over time. This assumption is unlikely to hold, however, especially during major devastating events like COVID-19. Therefore, the piecewise linear panel data model was applied in this study. As a result of empirical analysis, pre- and post-COVID differences were seen between different geographical regions. Based on analysis conducted for three distinct geographical regions with piecewise linear models, it was determined that the piecewise random effects model was appropriate for European and Central Asian countries, the piecewise FGLS model for Latin American and Caribbean countries, and the piecewise linear GMM model for South Asian countries. According to the results, there are many variables that affect happiness, which vary according to different geographical conditions and societies with different cultural values.
- 25 May 2023
- Published by MDPI AG in Sustainability
- Vol. 15 (11) , 8569
Open AccessExamination of the Effects of COVID-19 on Happiness in Different Geographical Regions with Piecewise Linear Panel Data Models
The COVID-19 pandemic has recently caused the loss of millions of lives, and billions of others have been deeply affected. This crisis has changed the way people live, think about life, and perceive happiness. The aim of this study is to reveal differences between geographical regions by investigating the effect of the happiness variable on different countries during the international COVID-19 pandemic. The primary purpose is to demonstrate how such a pandemic may affect different countries in terms of happiness at the individual level and to identify possible strategies for the future. With this aim, both static and dynamic panel data models were used while applying fixed effects, random effects, and the generalized method of moments (GMM). A basic assumption in panel data models is that the coefficients do not change over time. This assumption is unlikely to hold, however, especially during major devastating events like COVID-19. Therefore, the piecewise linear panel data model was applied in this study. As a result of empirical analysis, pre- and post-COVID differences were seen between different geographical regions. Based on analysis conducted for three distinct geographical regions with piecewise linear models, it was determined that the piecewise random effects model was appropriate for European and Central Asian countries, the piecewise FGLS model for Latin American and Caribbean countries, and the piecewise linear GMM model for South Asian countries. According to the results, there are many variables that affect happiness, which vary according to different geographical conditions and societies with different cultural values.
- Open Access
Path Analysis of the Relationships between the Eruption Time of the First Primary Teeth and Various Factors in Twins
The timing of primary tooth eruption is critical for children’s health planning and the diagnosis of specific growth disorders. The purpose of this study is to assess the relationship between twin pairs’ birth weight, gestational age, and gender, which are indicators of prenatal factors; breast-feeding duration, which is an indicator of postnatal factors; type of delivery, which is an indicator of maternal as well as genetic factors; and age of the primary tooth. Twin children aged from 3 to 15 years who applied to the clinic for the first dental examination constituted the sample group. In this twin study, 59 monozygotic (MZ) twin pairs and 143 dizygotic (DZ) twin pairs were included. Genetic (MZ vs. DZ), maternal (type of delivery, gestational age), perinatal (birth weight, gender), and postnatal (duration of breastfeeding) information was obtained, and effects on the children’s Eruption Timing of the First Primary Tooth (ETFPT) were examined. Statistical analysis was performed using the consistent partial least squares structural equation model (robust PLSc) technique. As birth weight increased, the age at first eruption became younger, but this change was different between MZ and DZ twins (p < 0.05). While the age at first tooth eruption was older in identical twins who were breastfed for the first 6 months, this increase was not observed in DZ twins. The mean of ETFPT was calculated as 7.31 months in MZ twins and 6.75 months in DZ twins. The effect of breastfeeding and birth weight on ETFPT may differ according to zygosity in twins. MZ twins may tend to take longer to experience the eruption of their first primary teeth.
Open AccessPath Analysis of the Relationships between the Eruption Time of the First Primary Teeth and Various Factors in Twins
The timing of primary tooth eruption is critical for children’s health planning and the diagnosis of specific growth disorders. The purpose of this study is to assess the relationship between twin pairs’ birth weight, gestational age, and gender, which are indicators of prenatal factors; breast-feeding duration, which is an indicator of postnatal factors; type of delivery, which is an indicator of maternal as well as genetic factors; and age of the primary tooth. Twin children aged from 3 to 15 years who applied to the clinic for the first dental examination constituted the sample group. In this twin study, 59 monozygotic (MZ) twin pairs and 143 dizygotic (DZ) twin pairs were included. Genetic (MZ vs. DZ), maternal (type of delivery, gestational age), perinatal (birth weight, gender), and postnatal (duration of breastfeeding) information was obtained, and effects on the children’s Eruption Timing of the First Primary Tooth (ETFPT) were examined. Statistical analysis was performed using the consistent partial least squares structural equation model (robust PLSc) technique. As birth weight increased, the age at first eruption became younger, but this change was different between MZ and DZ twins (p < 0.05). While the age at first tooth eruption was older in identical twins who were breastfed for the first 6 months, this increase was not observed in DZ twins. The mean of ETFPT was calculated as 7.31 months in MZ twins and 6.75 months in DZ twins. The effect of breastfeeding and birth weight on ETFPT may differ according to zygosity in twins. MZ twins may tend to take longer to experience the eruption of their first primary teeth.
- Open Access
Testing Nonlinearity with Rényi and Tsallis Mutual Information with an Application in the EKC Hypothesis
The nature of dependence between random variables has always been the subject of many statistical problems for over a century. Yet today, there is a great deal of research on this topic, especially focusing on the analysis of nonlinearity. Shannon mutual information has been considered to be the most comprehensive measure of dependence for evaluating total dependence, and several methods have been suggested for discerning the linear and nonlinear components of dependence between two variables. We, in this study, propose employing the Rényi and Tsallis mutual information measures for measuring total dependence because of their parametric nature. We first use a residual analysis in order to remove linear dependence between the variables, and then we compare the Rényi and Tsallis mutual information measures of the original data with that the lacking linear component to determine the degree of nonlinearity. A comparison against the values of the Shannon mutual information measure is also provided. Finally, we apply our method to the environmental Kuznets curve (EKC) and demonstrate the validity of the EKC hypothesis for Eastern Asian and Asia-Pacific countries.
Open AccessTesting Nonlinearity with Rényi and Tsallis Mutual Information with an Application in the EKC Hypothesis
The nature of dependence between random variables has always been the subject of many statistical problems for over a century. Yet today, there is a great deal of research on this topic, especially focusing on the analysis of nonlinearity. Shannon mutual information has been considered to be the most comprehensive measure of dependence for evaluating total dependence, and several methods have been suggested for discerning the linear and nonlinear components of dependence between two variables. We, in this study, propose employing the Rényi and Tsallis mutual information measures for measuring total dependence because of their parametric nature. We first use a residual analysis in order to remove linear dependence between the variables, and then we compare the Rényi and Tsallis mutual information measures of the original data with that the lacking linear component to determine the degree of nonlinearity. A comparison against the values of the Shannon mutual information measure is also provided. Finally, we apply our method to the environmental Kuznets curve (EKC) and demonstrate the validity of the EKC hypothesis for Eastern Asian and Asia-Pacific countries.
- Open Access
Measures of Qualitative Variation in the Case of Maximum Entropy
Asymptotic behavior of qualitative variation statistics, including entropy measures, can be modeled well by normal distributions. In this study, we test the normality of various qualitative variation measures in general. We find that almost all indices tend to normality as the sample size increases, and they are highly correlated. However, for all of these qualitative variation statistics, maximum uncertainty is a serious factor that prevents normality. Among these, we study the properties of two qualitative variation statistics; VarNC and StDev statistics in the case of maximum uncertainty, since these two statistics show lower sampling variability and utilize all sample information. We derive probability distribution functions of these statistics and prove that they are consistent. We also discuss the relationship between VarNC and the normalized form of Tsallis (α = 2) entropy in the case of maximum uncertainty.
Open AccessMeasures of Qualitative Variation in the Case of Maximum Entropy
Asymptotic behavior of qualitative variation statistics, including entropy measures, can be modeled well by normal distributions. In this study, we test the normality of various qualitative variation measures in general. We find that almost all indices tend to normality as the sample size increases, and they are highly correlated. However, for all of these qualitative variation statistics, maximum uncertainty is a serious factor that prevents normality. Among these, we study the properties of two qualitative variation statistics; VarNC and StDev statistics in the case of maximum uncertainty, since these two statistics show lower sampling variability and utilize all sample information. We derive probability distribution functions of these statistics and prove that they are consistent. We also discuss the relationship between VarNC and the normalized form of Tsallis (α = 2) entropy in the case of maximum uncertainty.