DSpace Arşivi :: by Yazar "Akkoyun, Serkan" değerine göre listeleniyor

Yazar "Akkoyun, Serkan" seçeneğine göre listele

Listeleniyor 1 - 20 / 47

Advanced predictive modelling of electric quadrupole transitions in even-even nuclei using various machine learning approaches
(Elsevier, 2025) Berbache, Sihem; Akkoyun, Serkan; Ali, Ahmed H.; Kartal, Sebahattin
Empirical predictions of electric quadrupole transition probabilities, B (E2; 0*-> 2*), in even-even nuclei, are among the principles needed to solve the nuclear structure and collective behaviour. In this study, nine different ML algorithms, gradient boosting machine (GBM), random forest (RF), convolutional neural network (CNN), k-nearest neighbour (KNN), CatBoost, extreme gradient boosting (XGBoost), neural network (NN), support vector machine (SVM) and multiple linear regression (MLR), are evaluated as a different data-driven solution for the prediction of B(E2) values. The outcomes show that ensemble models, in particular GBMs, RF, and XGBoost, provide vastly improved predictive capabilities and generalizing influence while creating strong correlations to experimental data with small prediction errors. On the other hand, deep learning models such as CNN and NN is prone to overfitting, while simpler ones such as MLR and KNN fail to capture the non-linear relationships inherent in nuclear data. The findings underscore the promise of ensemble ML tools for nuclear physics in a scalable, accurate approach for predicting transition probabilities.
alpha-decay half-life calculations of superheayy nuclei using artificial neural networks
(IOP PUBLISHING LTD, 2014) Bayram, Tuncay; Akkoyun, Serkan; Kara, S. Okan; Vagenas, EC; Vlachos, DS
Investigations of superheavy elements (SHE) have received much attention in the last two decades, due to the successful syntheses of SHE. In particular, alpha-decay of SHEs has a great importance because most synthesized SHE have alpha-decay and the experimentalists have evaluated the theoretical predictions of the alpha-decay half-life during the experimental design. Because of this, the correct prediction of alpha-decay half-life is important to investigate superheavy nuclei as well as heavy nuclei. In this work, artificial neural networks (ANN) have been employed on experimental alpha-decay half-lives of superheavy nuclei. Statistical modeling of alpha-decay half-life of superheavy nuclei have been found as to be successful.
An analysis of E(5) shape phase transitions in Cr isotopes with covariant density functional theory
(IOP PUBLISHING LTD, 2013) Bayram, Tuncay; Akkoyun, Serkan
Constrained relativistic mean field theory (RMF) has been employed for analysis of the shape phase transitions of even-even Cr52-66 isotopes. The systematic investigation of ground-state shape evolution between spherical U(5) and gamma-unstable O(6) for these nuclei has been carried out by using the potential energy curves (PECs) obtained from the effective interactions NL3*, TM1, PK1 and DD-ME2. The calculated PECs have indicated that Cr-58 can be a candidate for the critical-point nucleus with E(5) symmetry. A similar conclusion is also drawn from the calculated single-particle spectra of Cr-58 and is supported by the experimental data via observed ratios of the excitation energies. Furthermore, the gamma-independent character of Cr-58 has been pointed out by using its binding energy map in the beta-gamma plane obtained from the triaxial RMF calculations, in agreement with the condition of E(5) symmetry.
An approach to adjustment of relativistic mean field model parameters
(E D P SCIENCES, 2017) Bayram, Tuncay; Akkoyun, Serkan; Plompen, A; Hambsch, FJ; Schillebeeckx, P; Mondelaers, W; Heyse, J; Kopecky, S; Siegler, P; Oberstedt, S
The Relativistic Mean Field (RMF) model with a small number of adjusted parameters is powerful tool for correct predictions of various ground-state nuclear properties of nuclei. Its success for describing nuclear properties of nuclei is directly related with adjustment of its parameters by using experimental data. In the present study, the Artificial Neural Network (ANN) method which mimics brain functionality has been employed for improvement of the RMF model parameters. In particular, the understanding capability of the ANN method for relations between the RMF model parameters and their predictions for binding energies (BEs) of Ni-58 and Pb-208 have been found in agreement with the literature values.
An approximation to the cross sections of Z (l) boson production at CLIC by using neural networks
(VERSITA, 2013) Akkoyun, Serkan; Kara, Seyit O.
In this work, the possible dynamics associated with leptophilic Z (l) boson at CLIC (Compact Linear Collider) have been investigated by using artificial neural networks (ANNs). These hypotetic massive boson Z (l) have been shown through the process e (+) e (-)-> A mu(+)A mu(-). Furthermore, the invariant mass distributions for final muons have been consistently predicted by using ANN. For these highly non-linear data, we have constructed consistent empirical physical formulas (EPFs) by appropriate feed-forward ANN. These ANNEPFs can be used to derive further physical functions which could be relevant to studying Z (l) .
Calculating some nuclear properties of chromium isotopes in the shell model
(Indian Assoc Cultivation Science, 2024) Ali, Ahmed H.; Abbasi, Akbar; Akkoyun, Serkan; Korna, A. H.; Hossain, I.; Alshammari, H.; Zakaly, Hesham M. H.
The present study provides an in-depth theoretical examination of the shell model for a range of even-even Chromium (Z = 24) isotopes, encompassing neutron numbers both 22 and 36. The shell model calculations relied on assumptions about the disregarded core-polarization effects and the utilization of effective charges. We performed extensive theoretical calculations to determine the probability of reduced electric quadrupole transition, B(E2;0g.s+ -> 2+), the intrinsic quadrupole moments (Q0), the deformation parameters (beta 2,delta), and the inclusion of effective interactions such as fpd6, fpv, fpbm, and kb3. Using the NuShellX@MSU algorithm, the one-body density matrix elements (OBDM) were computed for these isotopes. Various effective charges were utilized in these computations, including NU-E effective charges obtained from the Nushellx@MSU software, ST-E standard effective charges, and BM-E effective charges calculated using Bohr and Mottelson's method. Comparative analysis was conducted between the theoretical values of transition rate B(E2), intrinsic quadrupole moments, deformation parameters and the available experimental data. The gained theoretical conclusions were subsequently contrasted with prior experimental data, which had similarly demonstrated the collapse of the magical property of the Cr isotope. The intrinsic quadrupole moment was optimal when employing the kb3 interaction, but the deformation parameter appeared optimal when using two interactions, fpbm and kb3. Furthermore, it has been demonstrated that the magical characteristic of the 52Cr (N = 28) isotope undergoes collapse.
Consistent empirical physical formula construction for recoil energy distribution in HPGe detectors by using artificial neural networks
(PERGAMON-ELSEVIER SCIENCE LTD, 2012) Akkoyun, Serkan; Yildiz, Nihat
The gamma-ray tracking technique is a highly efficient detection method in experimental nuclear structure physics. On the basis of this method, two gamma-ray tracking arrays, AGATA in Europe and GRETA in the USA, are currently being tested. The interactions of neutrons in these detectors lead to an unwanted background in the gamma-ray spectra. Thus, the interaction points of neutrons in these detectors have to be determined in the gamma-ray tracking process in order to improve photo-peak efficiencies and peak-to-total ratios of the gamma-ray peaks. In this paper, the recoil energy distributions of germanium nuclei due to inelastic scatterings of 1-5 MeV neutrons were first obtained by simulation experiments. Secondly, as a novel approach, for these highly nonlinear detector responses of recoiling germanium nuclei, consistent empirical physical formulas (EPFs) were constructed by appropriate feedforward neural networks (LFNNs). The LFNN-EPFs are of explicit mathematical functional form. Therefore, the LFNN-EPFs can be used to derive further physical functions which could be potentially relevant for the determination of neutron interactions in gamma-ray tracking process. (c) 2012 Elsevier Ltd. All rights reserved.
Consistent neural network empirical physical formula constructions for nonlinear scattering intensities of dye-doped nematic liquid crystals with ultraviolet pump laser-driven Fredericksz threshold shifts
(ELSEVIER GMBH, URBAN & FISCHER VERLAG, 2018) Polat, Omer; Yildiz, Nihat; Akkoyun, Serkan
Intrinsic high nonlinearity in experimentally measured laser scattering intensities poses significant difficulties in analyzing various molecular and optical properties of nematic liquid crystals (NLCs). In this respect, as we theoretically proved in a previous paper, universal nonlinear function approximator layered feedforward neural network (LFNN) can be applied to construct consistent empirical physical formulas (EPFs) for nonlinear physical phenomena. The novelty of this paper is that, by using our previous conference paper data (literature data or simply data for short) for He-Ne probe laser illumination nonlinear scattering intensities of dye-doped NLCs with ultraviolet pump laser-driven Fredericksz threshold (FT) shifts, we constructed definitive LFNN-EPFs for these illumination intensities of nonlinear scattering exhibiting FT shifts. The dyes used in the literature data were methyl red (MR) azo and disperse red (DR) anthraquinone. The LFNN-EPFs fitted the data very well. Moreover, magnificent LFNN test set forecastings over previously unseen data confirmed the consistent LFNN-EPFs inferences of the intensities of scattering. The LFNN-EPFs properly extracted the FT threshold shifts, as well as revealing the intensity dependencies on the kind of dye used. We, therefore, conclude the LFNN consistently infers nonlinear physical laws governing the NLC scattering data. Provided that sufficient scattering intensity data is available, these nonlinear physical laws embedded in LFNN-EPFs may potentially be useful for investigating various NLC molecular structure parameters in molecular nonlinear optics domain. This knowledge may be applicable in developing new optical materials. (C) 2017 Elsevier GmbH. All rights reserved.
Construction of consistent neural network empirical physical formulas for detector counts in neutron exit channel selection
(ELSEVIER SCI LTD, 2013) Akkoyun, Serkan; Yildiz, Nihat
Proper selection of neutron exit channels following heavy-ion reactions is important in nuclear structure physics. A knowledge of detector counts versus number of neutron interaction points per event can be useful in this selection. In this paper, we constructed layered feedforward neural networks (LFNNs) consistent empirical physical formulas (EPFs) to estimate the detector counts versus number of neutron interaction points per event. The LFNN-EPFs are of explicit mathematical functional form. Therefore, by various suitable operations of mathematical analysis, these LFNN-EPFs can be used to derivate further physical functions which might be potentially relevant to neutron exit channel selection. (C) 2013 Elsevier Ltd. All rights reserved.
Determination of Photonuclear Reaction Cross-Sections on Stable P-shell Nuclei by Using Deep Neural Networks
(Springer, 2023) Akkoyun, Serkan; Kaya, Huseyin; Seker, Abdulkadir; Yesilyurt, Saliha
Photonuclear reactions are widely used in investigations of nuclear structure. Thus, the determination of the cross-sections are essential for the experimental studies. In the present work, (gamma, n) photonuclear reaction cross-sections for stable p-shell nuclei have been estimated by using the neural network method. The main purpose of this study is to find neural network structures that give the best estimations for the cross-sections, and to compare them with the available data. These comparisons indicate the deep neural network structures that are convenient for this task. Through this procedure, we have found that the shallow NN models, tanh activation function is better than the ReLU. However, as our models become deeper, the difference between tanh and ReLU decreases considerably. In this context, we think that the crucial hyperparameters are the size of the hidden layer and neuron numbers of each layer.
Energy level and half-life determinations from photonuclear reaction on Ga target
(WORLD SCIENTIFIC PUBL CO PTE LTD, 2016) Akkoyun, Serkan; Bayram, Tuncay; Dulger, Fatih; Dapo, Haris; Boztosun, Ismail
Photonuclear reactions are important tools in the understanding of the nucleus. These reactions are also interesting for realizing the element creation processes in stellar environment. The use of bremsstrahlung photons generated from clinic linear accelerator is practical for performing these type of reactions. In this study, the bremsstrahlung photons with endpoint energy of 18 MeV have been used for activating gallium target material. After irradiation, the transition energies and half-lives associated with the decay of Ga-68, Ga-70 and Ga-72 isotopes have been determined The values obtained for half-life of Ga-68, Ga-70 and Ga-72 isotopes are 67.5 +/- 0.9 min, 21.1 +/- 0.9 min and 13.8 +/- 0.4 h, respectively. It has been seen that the values are consistent with the present literature values. In addition, the new measurements of gamma-ray energies for transition energies have been obtained comparable to the literature values with good uncertainties.
Estimation of fission barrier heights for even-even superheavy nuclei using machine learning approaches
(Iop Publishing Ltd, 2023) Yesilkanat, Cafer Mert; Akkoyun, Serkan
With the fission barrier height information, the survival probabilities of super-heavy nuclei can also be reached. Therefore, it is important to have accurate knowledge of fission barriers, for example, the discovery of super-heavy nuclei in the stability island in the super-heavy nuclei region. In this study, five machine learning techniques, Cubist model, Random Forest, support vector regression, extreme gradient boosting and artificial neural network were used to accurately predict the fission barriers of 330 even-even super-heavy nuclei in the region 140 <= N <= 216 with proton numbers between 92 and 120. The obtained results were compared both among themselves and with other theoretical model calculation estimates and experimental results. According to the results obtained, it was concluded that the Cubist model, support vector regression and extreme gradient boosting methods generally gave better results and could be a better tool for estimating fission barrier heights.
Estimation of fusion reaction cross-sections by artificial neural networks
(Elsevier, 2020) Akkoyun, Serkan
Accurate determination of total fusion and fusion-evaporation reaction cross-sections is an important task in experimental nuclear physics studies. In this study, we estimated the total fusion cross-sections and, as an example, one of the particular channels (2n) cross-sections for different reactions by using artificial neural network (ANN) methods. The root mean square errors for fusion reaction were obtained as 18.5 and 110.4 mb for the training and test data, which correspond to 1.8% and 10.5% deviations from the experimental cross-section values, respectively. These values for the 2n channel are 0.3% for training and 13.3% for test data of ANN. The deviations are mostly lower than the cross-section values from a commonly used theoretical calculation code. The results indicate that ANN methods might be a possible candidate tool for the estimation of cross-sections for fusion and fusion-evaporation reactions.
Estimation of the femur length from its proximal measurements in Anatolian Caucasians by artificial neural networks
(TAYLOR & FRANCIS LTD, 2016) Otag, Ilhan; Otag, Aynur; Akkoyun, Serkan; Cimen, Mehmet
Femora are a well preserved section of the skeleton after death. Therefore, they are commonly used in the field of forensic sciences, physical anthropology and anatomy. In addition, femur morphometry is helpful in finding sex or side (left or right) differences. The femur also shows characteristics of certain populations. Femur length is important for calculation of individual stature. In this study, the artificial neural network method was used to estimate femur length. In total, 230 femora exemplar were used. The three input parameters of the method were the distance between trochanter major top point and trochanter minor bottom point, the diameter of caput femoris and the diameter of collum femoris. By using these parameters, the artificial neural network estimation on femur length was performed. The results show that the method is capable of performing this estimation. In addition, sex discrimination was performed and achieved with 82% accuracy. As well as the identification of sex or side differences, morphometry of the proximal femur is necessary and important for surgical procedures.
Estimation of the future fracture epidemiology in the patients applying to the emergency department with long short time memory method
(2020) Pazarcı, Özhan; Torun, Yunis; Akkoyun, Serkan
Operation rooms, human resources and equipment planning are essential for increasing theeffectiveness of diagnostic and treatment methods in line with the needs of emergency cases.In this study, 151822 patients admitted to the emergency department (ED) within 3 yearswere examined in three categories including gender, fracture sites and causes of fracture.However, fracture cases were treated as time series and Long Short Time Memory (LSTM)method was used to estimate the number of future fracture cases. In the learning phase, thenumber of monthly cases in the next 6 months was estimated using 30-month case numbers.The Root Mean Square Error (RMSE), Mean Absolute Error (MAE) and Mean relative Error(MRE) values of the error rate between the estimated and actual number of cases were given.
Estimations of beta-decay energies through the nuclidic chart by using neural network
(PERGAMON-ELSEVIER SCIENCE LTD, 2014) Akkoyun, Serkan; Bayram, Tuncay; Turker, Tugba
One of the main characteristics of unstable nuclei is beta-decay energy (Q(beta)). It is determined by different methods such as beta endpoint measurements, counting in coincidence with annihilation radiation, electron capture (EC)/beta(+) ratio method, method of gamma absorption with X-ray coincidence. Beta-decay energy is a roughly linear function of atomic and mass numbers. Due to the fact that artificial neural network (ANN) is sufficient for nonlinear function approximation, in this study by using the nuclear masses from Hartree-Fock-BCS method, Q(beta) values have been obtained by ANN. It is seen that the estimations of the ANN are consistent with the calculated data within some deviation. (C) 2013 Elsevier Ltd. All rights reserved.
Estimations of Cross-Sections for Photonuclear Reaction on Calcium Isotopes by Artificial Neural Networks
(2020) Akkoyun, Serkan; Kaya, Hüseyin
The nuclear reaction induced by photon is one of the important tools in the investigation of atomic nuclei. In the reaction, a target material is bombarded by photons with high-energies in the range of gamma-ray energy range. In the bombarding process, the photons can statistically be absorbed by a nucleus in the target material. Then the excited nucleus can decay by emitting proton, neutron, alpha and light particles or photons. By performing photonuclear reaction on the target, it can be easily investigated low-lying excited states of the nuclei. In the present work, (?, n) photonuclear reaction cross-sections on different calcium isotopes have been estimated by using artificial neural network method. The method is a mathematical model that mimics the brain functionality of the creatures. The correlation coefficient values of the method for both training and test phases being 0.99 indicate that the method is very suitable for this purpose.Keywords: Photonuclear reaction, cross-section, calcium, artificial neural network
Estimations of fission barrier heights for Ra, Ac, Rf and Db nuclei by neural networks
(WORLD SCIENTIFIC PUBL CO PTE LTD, 2014) Akkoyun, Serkan; Bayram, Tuncay
Accurate information about the fission barrier is important for studying of the fission process. Fission barrier is needed for discovering the island of stability in superheavy region and searching of the superheavy elements. Furthermore, the astrophysical r-process is closely related to the fission barrier of the neutron-rich nuclei. In this study, by using artificial neural network (ANN) method, we have estimated the fission barrier heights of the Rf, Db, Ra and Ac nuclei covering 230 isotopes. For inner barrier calculation, we have used Rf and Db nuclei and the barrier heights have been determined between nearly 1 MeV and 7 MeV. The related mean square error value has been obtained as 0.108 MeV. For outer barrier calculation, we have used Ra and Ac nuclei and the heights have been determined between nearly 8 MeV and 28 MeV. The related mean square error has been obtained as 0.407. The results of this study indicate that ANN is capable for the estimations of inner and outer fission barrier heights.
Excited Energy Spectra for He and Be Isotopes from Different Shell Model Interactions
(Ankara University, 2021) Bektemir, İlker Ahmet; Akkoyun, Serkan
One of the common methods used to investigate the nuclear structures of atomic nuclei is the nuclear shell model. Like the arrangement of atomic electrons in orbitals, in the nuclear shell model, protons and neutrons are thought to be placed in orbits within the nucleus in accordance with the Pauli exclusion principle. These orbitals group together to form shells, which if all possible levels in a shell are filled, the shell is said to be closed. Atomic nuclei with closed shells are very stable and in nuclear shell model calculations, the valence nucleons more than these nuclei are included in the calculations. In this study, the nuclear shell model was used to investigate the nuclear structures of He and Be isotopes. 4He nucleus have been taken as a closed-shell core nucleus. Considering p3/2 and p1/2 as single-particle orbits in model space, different two-body interactions are used between valence nucleons. The results from the interactions giving better results were compared with each other and with the available experimental values.
Generation of fusion and fusion-evaporation reaction cross-sections by two-step machine learning methods
(Elsevier, 2024) Akkoyun, Serkan; Yesilkanat, Cafer Mert; Bayram, Tuncay
In order to obtain cross-sections of heavy-ion fusion and fusion-evaporation reactions, artificial neural networks, cubist, random forest, support vector regression, extreme gradient boosting, and multiple linear regression machine learning approaches were used separately in this study. The outcomes from these different methods that are obtained from the training carried out with the existing experimental data in the literature were compared. Furthermore, it has been observed that a two-step process yielded better results for determining the heavy ion reaction cross-sections, after first estimating which approach would be better for which reaction. In this manner, the method for which the cross-section needs to be calculated is determined by the machine learning classification application, and predictions can be made using the machine learning regression application with the determined method. It has been concluded that the obtained results are in harmony with the experimental data and that the methods can be used safely. The obtained results are published on a web page that allows for online calculation of heavy-ion fusion and fusion-evaporation reaction cross-sections.

Yazar "Akkoyun, Serkan" seçeneğine göre listele

Sayfa Başına Sonuç

Sıralama seçenekleri