Yazar "Oztemiz, Furkan" seçeneğine göre listele
Listeleniyor 1 - 11 / 11
Sayfa Başına Sonuç
Sıralama seçenekleri
Öğe A greedy approach to solve maximum independent set problem: Differential Malatya independent set algorithm(Elsevier - Division Reed Elsevier India Pvt Ltd, 2025) Oztemiz, FurkanIn this study, a method has been developed for solving the maximum independent set problem, which is one of the significant problems in graph theory. The maximum independent set problem is NP-hard for all types of graphs. The proposed method features a robust and greedy approach that produces results in polynomial time for all graph types. The proposed method is named the Differential Malatya Independent Set Algorithm (DMISA). The presented method also provides solutions to the minimum vertex cover and maximum clique problems, which are directly related to the independent set. The DMISA algorithm consists of two sub-algorithms. The first algorithm is the Differential Malatya Centrality Algorithm (DMCA), a centrality algorithm that calculates centrality values, providing prioritization in the selection of independent members. The second algorithm uses the DMCA value to select the independent set and vertex cover members in the graphs. In this study, the DMISA analytical proof has been applied to graphs with known solutions that can be solved in polynomial time. To emphasize the success of the algorithm, test operations have been conducted on various types of graphs. The conducted tests included 40 lattices, 40 bipartite, 24 multipartite, 32 social, and random graphs. The analysis results showed that DMISA produced optimal results in lattice, bipartite, and complete multipartite graphs, while it produced generally non-optimal results for randomly generated and social graphs. Additionally, DMISA is compared with MIS methods in popular graph libraries and 7 different MIS methods. In summary, DMISA produces a larger solution than standard greedy algorithms in experiments.Öğe An Effective Method for Determining Node Dominance Values: Malatya Centrality Algorithm(Ieee, 2024) Oztemiz, Furkan; Yakut, SelmanCentrality is a frequently used metric in graph theory that identifies the dominance values of nodes. In this study, the Malatya Centrality Algorithm, which approaches graph centrality from a new perspective, has been proposed. To calculate the centrality of a node in the graph, the ratio of the node's degree to the degrees of its neighboring nodes was computed and these ratios were summed. The results of the proposed algorithm were compared with established centrality algorithms in the literature and tested on well-known social networks such as Zachary's karate club, the dolphin social network, and the zebra network. The tests and analyses conducted demonstrate that the Malatya Centrality Algorithm is an effective and successful centrality algorithm.Öğe Analysis of Traffic Flow Profiles in a Transportation Network Using Graph Theory Methods(Institute of Electrical and Electronics Engineers Inc., 2024) Caglar, Akin; Oztemiz, FurkanThis study aims to identify critical intersections and green wave corridors in an urban transportation network by analyzing vehicle flows. The dataset used consists of Bluetooth data collected from 56 intersections in the city of Malatya. The transportation network and dataset are uniquely prepared for this study. For the centrality analysis of intersections, PageRank, Closeness, and Betweenness centrality methods were applied. To identify green wave corridors, Walktrap and Optimal community detection methods were utilized. The results of both applications are presented visually and numerically in the study. The analysis and visualization phases were carried out using the R programming language with the igraph package. The findings will serve as a crucial decision support system for improving urban traffic planning and identifying congested intersections. © 2024 IEEE.Öğe Communication between Interactive Virtual Characters and AI-Powered Language Learning Application(Institute of Electrical and Electronics Engineers Inc., 2025) Beytekin, Musa; Oztemiz, FurkanArtificial Intelligence(AI)-based technologies offerthe potential to transform language learning processes. How ever, there is a notablegap in the current literature regardinglanguage learning systems where users simultaneously interactwith multipleAI characters. This research conducts a comprehensive literature review analyzing existing conversational AIs,LangChain/LangGraph frameworks, and multi-character interaction systems. To address these gaps, a hybrid language learning platform development approach is proposed where userscan engage in context-sensitiveand consistent dialogues with multiple artificial intelligence characters simultaneously. In the experimental part, emotion analysis performances of five different large language models were comparatively evaluated foroptimal model selection. Test results revealed that O1 seriesmodels showed superior performance especially in negative-emotions (95-98%), while all models struggled with complexemotional state analysis (54-61%). The main contribution isaddressing the multi-character interaction gap in literatureand providing an empirical foundation for data-driven modelselection in hybrid systems. © 2025 IEEE.Öğe Determination of Efficient Green Wave Corridors in The Transportation Network with The Community Detection Method(Gazi Univ, 2024) Oztemiz, Furkan; Karci, AliSignaling systems play an important role in managing urban transport networks. Optimizing signaling systems significantly reduces traffic density in transport networks. One of the popular methods applied to increase the efficiency of the signaling system is the green wave application, which means the coordinated operation of the signaling systems. The green wave system prevents vehicles from being repeatedly caught in red light, reducing travel time, waiting time on the roads and carbon emissions of vehicles. The problem that will arise at this point is on which intersection points the green wave system will be applied. In this study, vehicle counting and signaling data of the city of Malatya were used and the transportation network data was converted into a weighted graph. By applying the walktrap community detection algorithm to the transportation network, the intersection points are grouped according to the vehicle transition similarities on them. It was applied to the green wave system physically for the intersection points in 2 different groups determined. The results show that there are significant increases in the number of vehicles passing per unit time in the regions where green waves are applied. This situation has resulted in a decrease in the number of vehicles waiting at red lights and significant reductions in carbon emissions from stationary vehicles into the atmosphere.Öğe Implementation of Applied Prediction with YSA for Data Groups with Association Rule(Ieee, 2017) Oztemiz, Furkan; Hamamei, Serder EthemNowadays, increasing data sizes have grown at incredible levels. Many firms want to interpret their produced data and reach the useful information. In this study, by making customer basket analysis of a company operating in the retail sector, to organize suitable campaigns for the customers has been aimed and sales amounts of campaign items have been predicted before the campaign. The association rules for the items purchased by the customers has been obtained by using the Apriori algorithm. Sales amounts of associated items have been predicted with Artificial Neural Networks (ANN). For prediction process with ANN, MATLAB-NNTOOL toolbox has been used. With these prediction process, sales amounts of the campaign items offered to the customers have been determined and an idea about the success of campaign success has been obtained. In the study, 34 months sales data has been considered.Öğe Kendall rank correlation analysis of Malatya Centrality Algorithm with well-known centrality measures(Yildiz Technical Univ, 2025) Yakut, Selman; Oztemiz, Furkan; Karci, AliThe concept of centrality is widely used in graph theory to determine the dominance of nodes within a graph. This concept is crucial for solving many real-life problems that are modeled using graphs. In this study, the effectiveness of a new approach, the Malatya Centrality Algorithm, for determining the centrality of nodes in a graph is examined. This algorithm provides effective solutions to problems in both graph theory and real-life applications. The centrality value in the Malatya Centrality Algorithm is calculated by summing the ratios of the degree of the relevant node to the degrees of its neighboring nodes. To demonstrate the effectiveness of the Malatya Centrality Algorithm, comparisons and analyses were conducted with well-known centrality algorithms in the literature. Various types of graphs, including random graphs, benchmark graphs, social network graphs, and lattice bipartite graphs, were used for these comparisons and analyses. Kendall rank correlation analysis and tests were performed on these different types of graphs for the Malatya Centrality Algorithm and the well-known centrality measures in the literature. The tests conducted on various graphs reveal the ranking of nodes based on their effectiveness. These rankings help identify nodes used in numerous problems. The tests and analyses demonstrate that the Malatya Centrality Algorithm produces results similar to those of established centrality algorithms in the literature and confirms its effectiveness across different types of graphs.Öğe KO: Modularity optimization in community detection(Springer London Ltd, 2023) Oztemiz, Furkan; Karci, AliMany algorithms have been developed to detect communities in networks. The success of these developed algorithms varies according to the types of networks. A community detection algorithm cannot always guarantee the best results on all networks. The most important reason for this is the approach algorithms follow when dividing any network into communities (sub-networks). The modularity of the network determines the quality of communities in networks. It is concluded that networks with high modularity values are divided into more successful communities (clusters, sub-networks). This study proposes a modularity optimization algorithm to increase clustering success in any network without being dependent on any community detection algorithm. The basic approach of the proposed algorithm is to transfer nodes at the community boundary to neighboring communities if they meet the specified conditions. The method called KO (Karci-Oztemiz) optimization algorithm maximizes the modularity value of any community detection algorithm in the best case, while it does not change the modularity value in the worst case. For the KO algorithm's test, in this study, Walktrap, Cluster Edge Betweenness, Label Propagation, Fast Greedy, and Leading Eigenvector community detection algorithms have been applied on three popular networks that were unweighted and undirected previously used in the literature. The community structures created by five community detection algorithms were optimized via the KO algorithm and the success of the proposed method was analyzed. When the results are examined, the modularity values of the community detection algorithms applied on the three different networks have increased at varying rates (0%,.,14.73%).Öğe Malatya Public Transportation Route Optimization via Ant Colony Algorithm(Ieee, 2018) Oztemiz, Furkan; Yeroglu, CelaleddinIncreasing population density causes traffic densities in city centers. In this study, Ant Colony Algorithm (ACO) was used to find solutions to the traffic problems in crowded cities and Malatya province was chosen as the application region. Need of reducing the traffic intensity in the city centers, has led to the idea that the central stop of public transportation vehicles should be moved. This situation reveals the problem of changing the routes of public transport. In this study, ACO algorithm was used to analyze the new routes in the most ideal way. It is aimed to realize minimum distance and minimum traffic density by solving this problem which is similar to the traveling salesman problem. In order to achieve minimum traffic intensity, the threshold pheromone amount is determined to direct multiple vehicles to alternative routes. The data used in the analysis belongs to the public transportation vehicles of the city of Malatya. A java based program was used to construct the datasets and to solve the problem.Öğe Performance Comparison of Physics Based Meta-Heuristic Optimization Algorithms(Ieee, 2018) Demirol, Doygun; Oztemiz, Furkan; Karci, AliThe optimization process is a process that aims to find the minimum or maximum point according to the objective function. Many different algorithms have been developed for optimization problems. While analytical methods are committed to finding the exact solution specific to their problem, heuristic methods are committed to finding the best solution to the larger set of problems. Mathematical models of the system and the objective function are needed to solve the problems. General purpose heuristic optimization algorithms are evaluated in eight different groups including physics, biology, social, herd, music, chemistry, sports and mathematics. In this study, Be about Water Cycle Algorithm, Electromagnetic Field Optimization, Big Bang Big Crunch, Gravitational Search Algorithm, Optics Inspired Optimization, the performance results of 5 different algorithms were compared for Sphere, Rastrigin, Rosenbrock, Griewank and Ackley test functions. In consequence of the number of stated population, size, run and iteration, after the minimum, maximum, standard deviation, and their mean values were established, their superiority to each other was determined.Öğe Using Clique to Identify Access Points with High-Density Coverage in Different Locations(Institute of Electrical and Electronics Engineers Inc., 2024) Oztemiz, FurkanThis study aims to identify sensor network devices located in areas of high connection density within a network. By doing so, it will be possible to determine regions with dense and more comprehensive connectivity. The dataset used in this study is modeled as a disc graph representing a university's campus network. In the disc graph, overlapping areas are depicted as edges in the original graph, with the number of these edges indicating the extent of connectivity and coverage. To identify these dense edges in the designed network, the maximum clique approach has been applied. The members of the maximum clique have been determined using the Malatya independent set algorithm. The study aims to prevent the overlap of maximum cliques in the same area, focusing on identifying maximum clique members that are not directly connected to each other. The resulting data provides critical insights into the hub points of the network and the optimal placement of main routers. © 2024 IEEE.
