Bilgisayar ve Bilişim Fakültesi Koleksiyonu

Permanent URI for this collectionhttps://hdl.handle.net/20.500.13091/10834

Browse

Now showing 1 - 12 of 12

Citation - Scopus: 1
Automatic Sleep Stage Classification for the Obstructive Sleep Apnea
(Trans Tech Publications Ltd, 2023) Özsen, Seral; Koca, Yasin; Tezel, Gülay Tezel; Solak, Fatma Zehra; Vatansev, Hulya; Kucukturk, Serkan
Automatic sleep scoring systems have been much more attention in the last decades. Whereas a wide variety of studies have been used in this subject area, the accuracies are still under acceptable limits to apply these methods to real-life data. One can find many high-accuracy studies in literature using a standard database but when it comes to using real data reaching such high performance is not straightforward. In this study, five distinct datasets were prepared using 124 persons including 93 unhealthy and 31 healthy persons. These datasets consist of time-, nonlinear-, welch-, discrete wavelet transform- and Hilbert-Huang transform features. By applying k-NN, Decision Trees, ANN, SVM, and Bagged Tree classifiers to these feature sets in various manners by using feature-selection highest classification accuracy was searched. The maximum classification accuracy was detected in the case of the Bagged Tree classifier as 95.06% with the use of 14 features among a total of 136 features. This accuracy is relatively high compared with the literature for a real-data application.
A Comparative Application Regarding the Effects of Traveling Salesman Problem on Logistics Costs
(2019) Dündar, Abdullah Oktay; Şahman, Mehmet Akif; Tekin, Mahmut; Kıran, Mustafa Servet
The necessity of transporting goods from production facilities to buyers requires every company to manage logistics. While the quantity of products ordered has been decreasing in recent years, the number of orders has been increasing. This situation leads to higher logistics costs and more attempts to control logistics costs by business managers. One way to decrease logistics costs is the optimization of traveled distances. The Traveling Salesman Problem (TSP) attempts to optimize travel distances by changing the order of the locations to be visited. By doing so, it reduces the logistics costs associated with travel distances. However, there are also some parameters of logistics costs that are not related to travel distances. This paper examines the effects of optimization results by TSP on logistics costs, using seven different methods to consider a real logistics problem, and comparing the results. Then it discusses the variation in logistics costs due to TSP.
Citation - Scopus: 4
A Comprehensive Study of Parameters Analysis for Galactic Swarm Optimization
(Ismail Saritas, 2021) Kaya, Ersin
The galactic swarm optimization algorithm is a metaheuristic approach inspired by the motion and behavior of stars and galaxies. It is a framework that can use basic metaheuristic search methods. The method, which has a two-phase structure, performs exploration in the first phase and exploitation in the second phase. GSO tries to find the best solution in the search space by repeating these two phases for the specified number of times. In this study, the analysis of maximum epoch number (EPmax), the number of iterations in the first phase (L1), and the number of iterations in the second phase (L2) parameters, which determine the exploration and exploitation balance in the GSO method, was performed. 15 different parameter sets consisting of different values of these three parameters were created. The methods with 15 different parameter sets were performed at 30 independent runs. The methods were analyzed using 26 benchmark functions. The functions are tested in 30, 60, and 100 dimensions. Detailed results of the analysis were presented in the study, and the results obtained were also evaluated statistically. © 2021, Ismail Saritas. All rights reserved.
Citation - Scopus: 4
Deep Learning-Based Brain Hemorrhage Detection in Ct Reports
(IOS Press BV, 2022) Bayrak, Gıyaseddin; Toprak, M. Şakir; Ganiz, Murat Can; Kodaz, Halife; Koç, Ural
Radiology reports can potentially be used to detect critical cases that need immediate attention from physicians. We focus on detecting Brain Hemorrhage from Computed Tomography (CT) reports. We train a deep learning classifier and observe the effect of using different pre-trained word representations along with domain-specific fine-tuning. We have several contributions. Firstly, we report the results of a large-scale classification model for brain hemorrhage detection from Turkish radiology reports. Second, we show the effect of fine-tuning pre-trained language models using domain-specific data on the performance. We conclude that deep learning models can be used for detecting brain Hemorrhage with reasonable accuracy and fine-tuning language models using domain-specific data to improve classification performance. © 2022 European Federation for Medical Informatics (EFMI) and IOS Press.
Citation - Scopus: 1
Feature Extraction Methods for Predicting the Prevalence of Heart Disease
(Springer Science and Business Media Deutschland GmbH, 2022) Ngong, I.C.; Baykan, Nurdan
This paper presents an automatic classification technique for the detection of cardiac arrhythmias from ECG signals. With cardiac arrhythmias being one of the leading causes of death in the world, accurate and early detection of beat abnormalities can significantly reduce mortality rates. ECG signals are vastly used by physicians for diagnosing heart problems and abnormalities as a result of its simplicity and non-invasive nature. The aim of this study is to determine the most accurate combination of feature extraction methods and SVM (Support Vector Machine) kernel classifier that will produce the best results on ECG signals obtained from the MIT-BIH Arrhythmia Database. SVM classifiers with four different kernels (linear, polynomial, radial basis, and sigmoid) were used to classify different features extracted from the four feature selection methods; Random Forests, XGBoost, Principal Component Analysis, and Convolutional Neural Networks. The CNN-SVM classifier produced the best results overall, with the polynomial kernel achieving the maximum accuracy of 99.2%, the best sensitivity 92.40% from the radial basis kernel, and best specificity of 98.92% from the linear kernel. The high classification accuracy obtained is comparable to or even better than other approaches in literature. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.
Citation - WoS: 2
Citation - Scopus: 3
Fusion and Cnn Based Classification of Liver Focal Lesions Using Magnetic Resonance Imaging Phases
(Yildiz Technical University, 2023) Ci̇han, M.; Uzbaş, B.; Ceylan, M.
The diagnosis and follow-up of focal liver lesions have an important place in radiology practice and in planning the treatment of patients. Lesions detected in the liver can be benign or malign. While benign lesions do not require any treatment, some treatments and surgical operations may be required for malign lesions. Magnetic resonance imaging provides some advantages over other imaging modalities in the detection and characterization of focal liver lesions with its superior soft tissue contrast. Additionally, different phases help make a clear diagnosis of different contrast agent retention properties in magnetic resonance imaging. This study aims to classify focal liver lesions based on convolutional neural networks by fusing magnetic resonance liver images obtained in pre-contrast, venous, arterial, and delayed phases. Magnetic resonance imaging data were obtained from Selcuk University, Faculty of Medicine, Department of Radiology in Turkey. The experiments were performed using 460 magnetic resonance images in four phases of 115 patients. Two experiments were conducted. Two-dimensional discrete wavelet transform was used to fuse the phases in both experiments. In the first experiment, the best model was determined using the original data, different number of convolution layers and different activation functions. In the second experiment, the best-found model was used. Additionally, the number of data was increased using data augmentation methods in this experiment. The results were compared with other state-of-the art methods and the superiority of the proposed method was proved. As a result of the classification, 96.66% accuracy, 86.67% sensitivity and 98.76% specificity rates were obtained. When the results are examined, CNN efficiency increases by fusing MR liver images taken in different phases. © 2021, Yıldız Technical University.
Citation - Scopus: 1
Gender Determination From Teeth Images Via Hybrid Feature Extraction Method
(SPRINGER INTERNATIONAL PUBLISHING AG, 2020) Uzbaş, Betül; Arslan, Ahmet; Kök, Hatice; Acılar, Ayşe Merve
Teeth are a significant resource for determining the features of an unknown person, and gender is one of the important pieces of demographic information. For this reason, gender analysis from teeth is a current topic of research. Previous literature on gender determination have generally used values obtained through manual measurements of the teeth, gingiva, and lip area. However, such methods require extra effort and time. Furthermore, since sexual dimorphism varies among populations, it is necessary to know the optimum values for each population. This study uses a hybrid feature extraction method and a Support Vector Machine (SVM) for gender determination from teeth images. The study group was composed of 60 Turkish individuals (30 female, 30 male) between the ages of 19 and 27. Features were automatically extracted from the intraoral images through a hybrid method that combines two-dimensional Discrete Wavelet Transformation (DWT) and Principle Component Analysis (PCA). Classification was performed from these features through SVM. The system can be easily used on any population and can perform fast and low-cost gender determination without requiring any extra effort.
A Modified Artificial Algae Algorithm for Large Scale Global Optimization Problems
(2018) Uymaz, Sait Ali; Koçer, Havva Gül
Optimization technology is used to accelerate decision-making processes and to increase the quality of decision making inmanagement and engineering problems. The development technology has made real world problems large and complex. Many optimizationmethods that proposed for solving large-scale global optimization (LSGO) problems suffer from the “curse of dimensionality”, whichimplies that their performance deteriorates quickly as the dimensionality of the search space increases. Therefore, more efficient and robustalgorithms are needed. When literature on large-scale optimization problems is examined, it is seen that algorithms with effective globalsearch ability have better results. For the purpose, in this paper Modified Artificial Algae Algorithm (MAAA) is proposed by modifyingoriginal version of Artificial Algae Algorithm (AAA) inspiring by Differential Evolution Algorithm (DE)’s mutation strategies. AAA andMAAA are compared with each other by operating with the first 10 benchmark functions of CEC2010 Special Session on Large ScaleGlobal Optimization. The results show that hybridization process that applied by updating an additional fourth dimension with mutationstrategies of DE after the helical motion of the AAA algorithm, contributes exploration phase and improves the AAA performance onLSGO.
A New Variable Ordering Method for the K2 Algorithm
(SPRINGER INTERNATIONAL PUBLISHING AG, 2020) Uzbaş, Betül; Arslan, Ahmet
K2 is an algorithm used for learning the structure of a Bayesian networks (BN). The performance of the K2 algorithm depends on the order of the variables. If the given ordering is not sufficient, the score of the network structure is found to be low. We proposed a new variable ordering method in order to find the hierarchy of the variables. The proposed method was compared with other methods by using synthetic and real-world data sets. Experimental results show that the proposed method is efficient in terms of both time and score.
A Novel Multi-Swarm Approach for Numeric Optimization
(2018) Babalık, Ahmet
In order to solve the numeric optimization problems, swarm-based meta-heuristic algorithms can be used as an alternative to solve optimization problems. Meta-heuristic algorithms do not guarantee finding the optimal solution but they produce acceptable solutions in a reasonable computation time. By depending on the nature of the problems and the structure of the meta-heuristic algorithms, different results are obtained by different algorithms, and none of the meta-heuristic algorithm could guarantee to find the optimal solution. Particle swarm optimization (PSO) and artificial bee colony (ABC) algorithms are well known meta-heuristic algorithms often used for solving numeric optimization problems. In this study, a novel multi-swarm approach based on PSO and ABC algorithms is suggested. The proposed multi-swarm approach includes PSO and ABC algorithms together and replacing the swarm which achieves better solutions than the other algorithm in a pre-defined migration period. By this migration, swarm always include better solutions concerned to the algorithm which achieves better results. While running PSO and ABC algorithms competitively, this migration ensures to utilize better solutions of both the solutions of PSO or ABC algorithms, and the convergence characteristic of each algorithm provides different approximation to the solution space. Thus, it is expected to obtain successful solutions and increasing the success rate at each migration cycle. The suggested approach has been tested on 14 well-known benchmark functions, and the results of the study are compared with the results in literature. The experimental results and comparisons show that the proposed approach is better than the other algorithms.
Citation - Scopus: 1
Segmentation-Based 3d Point Cloud Classification on a Large-Scale and Indoor Semantic Segmentation Dataset
(Springer Science and Business Media Deutschland GmbH, 2021) Sağlam, Ali; Baykan, Nurdan Akhan
This paper presents a segmentation-based classification technique for 3D point clouds. This technique is supervised and needs a ground-truth data for the training process. In this work, the Stanford Large-Scale 3D Indoor Spaces (S3DIS) dataset has been used for the classification of points with the segmentation pre-processing. The dataset consists of a huge amount of points and has semantic ground-truth segments (structures and objects). The main problem in this study is to classify raw points according to the predefined objects and structures. For this purpose, each semantic segment in the training part is segmented separately by a novel successful segmentation algorithm at first. The extracted features of each sub-segments resulted from the segmentation of the semantic segments in the training part are trained using the classifier, and a trained model is obtained. Finally, the raw data reserved for testing are segmented using the same segmentation parameters as used for training, and the result segments are classified using the trained model. The method is tested using two classifiers which are Support Vector Machine (SVM) and Random Forest (RF) with different segmentation parameters. The quantitative results show that RF gives a very useful classification output for such complicated data. © 2021, The Author(s), under exclusive license to Springer Nature Switzerland AG.
Citation - Scopus: 17
Tree-Seed Programming for Modelling of Turkey Electricity Energy Demand
(Ismail Saritas, 2022) Kıran, Mustafa Servet; Yunusova, P.
Tree-Seed algorithm, TSA for short, is a population-based metaheuristic optimization algorithm proposed for solving continuous optimization problems inspired by the relation between trees and their seeds in nature. The artificial agents in TSA are trees and seeds which correspond to possible solutions to the optimization problem, and the optimization procedure is executed by the interaction between trees and seeds. In this study, a programming version of this algorithm by using a crossover solution generation mechanism has been proposed. The proposed algorithm is called TSp and its performance has been investigated on two problems, one of them is symbolic regression benchmark functions and the other is the long-term energy estimation model of Turkey. Firstly, the continuous parts of TSA, which are initialization and solution generation mechanisms, have been modified to solve automatic programming problems. The solution representation is also modified to solve the problem addressed by the study. As a result of these modifications, TSp has been obtained and applied to symbolic regression problems for performance judgment, energy estimation problems for real-world application. The experimental results of TSp have been compared with those of Genetic Programming, it is concluded that TSp is better than the GP in solving energy estimation problems. © 2022, Ismail Saritas. All rights reserved.

Browse

Browsing Bilgisayar ve Bilişim Fakültesi Koleksiyonu by Scopus Q "Q4"