Yazar "Turker, Ilker" seçeneğine göre listele
Listeleniyor 1 - 16 / 16
Sayfa Başına Sonuç
Sıralama seçenekleri
Öğe Arithmetic success and gender-based characterization of brain connectivity across EEG bands(Elsevier Sci Ltd, 2021) Demir, Sait; Turker, IlkerWe provide a distinctive view for rest/task, gender and arithmetic success state for human brain concerning EEG-based functional brain networks. Utilizing coherence method to quantify phase synchronization between EEG nodes, signal activities are converted into graph representations. After a complex-theoretic approach is conducted, intelligent brains emerge as more connected ones under resting state. Male brain, featuring lower connection strength and efficiency under resting state, exhibits the ability to boost up connectivity under mental workload. On the other hand, arithmetic success correlates with high resting state connectivity for all EEG bands but dominantly for gamma band, while unsuccessful brains yield greater beta band assortativity behavior. Theta band associated with unconscious actions apparently exhibits greater connection weights for mental activity compared to resting. Contributions of EEG bands to diagnosing differences in rest/task, gender and arithmetic success states are detailed within the study. We also spot out which connection patterns are related with mental progressing, outlining that intelligent brains yield less inter-frontal and more frontal to central and frontal to parieto-occipital connections.Öğe Classification of cardiac disorders using weighted visibility graph features from ECG signals(Elsevier Sci Ltd, 2024) Kutluana, Gokhan; Turker, IlkerAs universal expressions to describe complex systems, graphs are increasingly preferred as a representation method in artificial intelligence. Visibility graphs enable converting time-series data into graph representations, inheriting some key properties of the series. This study investigates the representation capacity of visibility graphs for ECG signals using either the sequence of node weights or the diagonals of the adjacency matrices as feature sets, input to ResNet and Inception classifier models. This approach also reduces the high dimensionality of the original graph representation which features a size of data points squared. Experiments performed on the multi-labeled PTB-XL dataset indicate that the first 3 diagonals of the visibility graph as the feature set to the ResNet model provides superior classification results compared to the original signal, node weights from the visibility graph, or the combinations of these inputs. Having achieved a maximum AUC score of 93.46%, this approach also outperforms the previously recorded ECG classification results for the PTB-XL dataset.Öğe Connectogram - A graph-based time dependent representation for sounds(Elsevier Sci Ltd, 2022) Turker, Ilker; Aksu, SerkanThe proposed method contributes the time-series classification literature with a novel time-convexity based representation, which extends the current graph conversion approaches by introducing the time dimension, also introducing a colorful graph-generator approach. The representation capability of connectograms is tested in comparison with mel-spectrograms (mels) and MFCCs for an environmental sound classification task, as input to state-of-art transfer learning models. Results indicate that connectograms cannot compete with the best-performer mel-spectrogram representations in standalone format, however they significantly improve their classification performance in case they are combined as single layers of hybrid RGB representations. A combination of [mels + mels + connectogram] outperforms either sole representations or their combinations by 2-3%, with 96.46% classification accuracy for ResNet50 classifier model.(c) 2022 Elsevier Ltd. All rights reserved.Öğe Detailing the co-authorship networks in degree coupling, edge weight and academic age perspective(Pergamon-Elsevier Science Ltd, 2016) Turker, Ilker; Cavusoglu, AbdullahScientific collaboration networks are good resources for understanding self-organizing systems, reflecting the main generic properties like clustering, small-world and scale-free degree distribution. Beyond discovering the evolution of main parameters, we aimed to uncover the microscopic wiring properties in this study. We focused on the degree circumstances of pairing nodes together with degree differences, academic age differences and link weights. Analyzes are visualized by single distribution plots of the network parameters together with the 2D coupling characteristics of these parameters with a logarithmic colorbar as a third dimension, drawing visual perspective presenting who prefers connecting to whom during the network evolution. We showed that majority of the edges in the co-authorship network connects the nodes of comparable degrees and academic ages, featuring that strong collaboration activities occur between comparable academic careers. We also stated out that beyond the node degree distributions, power-law regimes are also observed in link weight and degree difference distributions. (C) 2016 Elsevier Ltd. All rights reserved.Öğe Generating clustered scale-free networks using Poisson based localization of edges(Elsevier, 2018) Turker, IlkerWe introduce a variety of network models using a Poisson-based edge localization strategy, which result in clustered scale-free topologies. We first verify the success of our localization strategy by realizing a variant of the well-known Watts-Strogatz model with an inverse approach, implying a small-world regime of rewiring from a random network through a regular one. We then apply the rewiring strategy to a pure Barabasi-Albert model and successfully achieve a small-world regime, with a limited capacity of scale-free property. To imitate the high clustering property of scale-free networks with higher accuracy, we adapted the Poisson-based wiring strategy to a growing network with the ingredients of both preferential attachment and local connectivity. To achieve the collocation of these properties, we used a routine of flattening the edges array, sorting it, and applying a mixing procedure to assemble both global connections with preferential attachment and local clusters. As a result, we achieved clustered scale-free networks with a computational fashion, diverging from the recent studies by following a simple but efficient approach. (C) 2018 Elsevier B.V. All rights reserved.Öğe A HIERARCHICAL VIEW OF A NATIONAL STOCK MARKET AS A COMPLEX NETWORK(Acad Economic Studies, 2017) Baydilli, Yusuf Yargi; Bayir, Safak; Turker, IlkerWe created a financial network for Borsa Istanbul 100 Index (BIST-100) which forms of N=100 stocks that bargained during T=2 years (2011-2013). We analyzed the market via minimum spanning tree (MST) and hierarchical tree (HT) by using filtered correlation matrix. While using hierarchical methods in order to investigate factors that affecting grouping of stocks, we have taken account the other statistical and data mining methods to examine success of stock correlation network concept for portfolio optimization, risk management and crisis analysis. We observed that financial stocks, especially Banks, are central position of the network and control information flow. Besides the sectoral and sub-sectoral behavior, corporations play role at grouping of stocks. Finally, this technique provided important tips for determining risky stocks in market.Öğe Is the world small enough? - A view from currencies(World Scientific Publ Co Pte Ltd, 2019) Baydilli, Yusuf Yargi; Turker, IlkerExchange rates are important indicators of the economic power of countries, directly affected by the international trading patterns and relations. Since almost every pair of countries in the globalized world are economically and financially related, exchange rates can be evaluated as nodes of a global financial network to make meaningful inferences. In this study, a financial network approach is conducted by evaluating the movements of the most traded 35 currencies against gold between years 2005 and 2017. Using graph theory and statistical methods, the analysis of economic relations between currencies is carried out, supported with geographical and cultural inferences. A risk map of currencies is generated through the portfolio optimization. Another approach of applying various threshold levels for correlations to determine connections between currencies is also employed. Results indicate that there exists a saddle point for correlation threshold as which results in a robust network topology that is highly modular and clustered, also dominantly displaying small-world and scale-free properties.Öğe Lung Disease Detection Using U-Net Feature Extractor Cascaded by Graph Convolutional Network(Mdpi, 2024) Rashid, Pshtiwan Qader; Turker, IlkerComputed tomography (CT) scans have recently emerged as a major technique for the fast diagnosis of lung diseases via image classification techniques. In this study, we propose a method for the diagnosis of COVID-19 disease with improved accuracy by utilizing graph convolutional networks (GCN) at various layer formations and kernel sizes to extract features from CT scan images. We apply a U-Net model to aid in segmentation and feature extraction. In contrast with previous research retrieving deep features from convolutional filters and pooling layers, which fail to fully consider the spatial connectivity of the nodes, we employ GCNs for classification and prediction to capture spatial connectivity patterns, which provides a significant association benefit. We handle the extracted deep features to form an adjacency matrix that contains a graph structure and pass it to a GCN along with the original image graph and the largest kernel graph. We combine these graphs to form one block of the graph input and then pass it through a GCN with an additional dropout layer to avoid overfitting. Our findings show that the suggested framework, called the feature-extracted graph convolutional network (FGCN), performs better in identifying lung diseases compared to recently proposed deep learning architectures that are not based on graph representations. The proposed model also outperforms a variety of transfer learning models commonly used for medical diagnosis tasks, highlighting the abstraction potential of the graph representation over traditional methods.Öğe A multilayer network analysis of hashtags in twitter via co-occurrence and semantic links(World Scientific Publ Co Pte Ltd, 2018) Turker, Ilker; Sulak, Eyub EkmelComplex network studies, as an interdisciplinary framework, span a large variety of subjects including social media. In social networks, several mechanisms generate miscellaneous structures like friendship networks, mention networks, tag networks, etc. Focusing on tag networks (namely, hashtags in twitter), we made a two-layer analysis of tag networks from a massive dataset of Twitter entries. The first layer is constructed by converting the co-occurrences of these tags in a single entry (tweet) into links, while the second layer is constructed converting the semantic relations of the tags into links. We observed that the universal properties of the real networks like small-world property, clustering and power-law distributions in various network parameters are also evident in the multilayer network of hashtags. Moreover, we outlined that co-occurrences of hashtags in tweets are mostly coupled with semantic relations, whereas a small number of semantically unrelated, therefore random links reduce node separation and network diameter in the co-occurrence network layer. Together with the degree distributions, the power-law consistencies of degree difference, edge weight and cosine similarity distributions in both layers are also appealing forms of Zipf's law evident in nature.Öğe Patterns of collaboration in four scientific disciplines of the Turkish collaboration network(Elsevier, 2014) Cavusoglu, Abdullah; Turker, IlkerScientific collaboration networks, as a prototype of complex evolving networks, are studied in many aspects of their structure and evolving characteristics. The organizing principles of these networks also vary in different scientific disciplines, demonstrating that each discipline has specific connecting rules. Retrieving the co-authorship data from the ISI Web of Science, we constructed networks of four disciplines (engineering, mathematics, physics and surgery) as a subset of the Turkish scientific collaboration network spanning 33 years' data, To provide a comparative perspective on the network topologies, we studied some statistical and topological properties such as the number of authors, degree distributions, authors per paper and papers per author histograms and distributions. These properties yield that the rapid growth of high education in Turkey (i.e. doubling of the number of universities and students within the last decade) had boosted the number of publications and increased the level of collaborations in the scientific collaboration networks. We showed the occurrence of Matthew effect in career longevity distributions, and also outlined the Heaps' law relation in the scaling of the collaborations as well. We outlined the prominent properties of each subset, while the similarities and deviations from the interdisciplinary networks are also evaluated. (C) 2014 Elsevier B.V. All rights reserved.Öğe Scientific collaboration network of Turkey(Pergamon-Elsevier Science Ltd, 2013) Cavusoglu, Abdullah; Turker, IlkerNetworking via co-authorship is an important area of research and used in many fields such as ranking of the universities/departments. Studying on the data supplied by the Web of Science, we constructed a structural database that defines the scientific collaboration network of the authors from Turkey, based on the publications between 1980 and 2010. To uncover the evolution and structure of this complex network by scientific means, we executed some empirical measurements. The Turkish scientific collaboration network is in an accelerating phase in growth, highly governed by the national policies aiming to develop a competitive higher education system in Turkey. As our results suggest the authors tend to make more number of collaborations in their studies over the years. The results also showed that, node separation of the network slightly converges about 4, consistent with the small world phenomenon. Together with this key indicator, the high clustering coefficient, (which is about 0.75) reveals that our network is strongly interconnected. Another quantity of major interest about such networks is, the degree distribution. It has a power-law tail that defines the network as scale-free. Along with the final values, the time evolutions of the above-mentioned parameters are presented in detail with this work. In a good agreement with the recent studies, our network yields some significant differences especially in growing rate, clustering properties and node separation. In contrast with the recent studies, we also showed that preferring to attach popular nodes result with being a more popular node in the future. (C) 2013 Elsevier Ltd. All rights reserved.Öğe The scientific studies on smart grid in selected European countries(E D P Sciences, 2017) Tan, Serhat Orkun; Turker, Ilker; Toku, TurkerSmart grid is a power system consisting of many transmission and distribution systems subjected to an automation which are efficient, reliable and coordinated with each other. As a nature friendly technology, Smart grid come into prominence due to the increasing energy consumption and limited renewable energy sources around the world. In the near future, the use of renewable energy sources is not expected to grow rapidly; but the transmission and distribution systems will be enhanced by Smart grid technologies. Considering these significant benefits, the studies have been increased on Smart grid technologies to meet the energy requirement in each country. Herewith, the aim of this study is to analyse the scientific studies in developed European countries such as Italy, Germany, United Kingdom, France and Spain to find out the increment rate of the importance devoted to the Smart grid technologies in academicals manner. The scientific researches on Smart grid are achieved from the Web of Science database and the statistical analysis have been made by utilizing proper SQL queries in combination with Excel Power Pivot for these countries. The correlation between the scientific studies on smart grid and the virtual smart grid applications are also outlined for each selected country.Öğe Solar radiation exergy and enviroeconomic analysis for Turkey(Inderscience Enterprises Ltd, 2017) Kurtgoz, Yusuf; Deniz, Emrah; Turker, IlkerThe knowledge of useful solar energy amount is important for designing solar energy systems. In this study, solar radiation exergy (SRE) values are calculated for the 81 cities in Turkey. The data are obtained from the General Directorate of Renewable Energy and the General Directorate of Meteorology. The approaches of Petela, Spanner and Jeter are used to calculate the exergy-to-energy ratios for determining the maximum utilisable solar radiation energy. The exergy-to-energy ratios, the SRE values, the locations having the highest solar exergy potential and the monthly, seasonal and annual total SRE maps are presented for Turkey. Finally, enviroeconomic analysis is performed and its results are presented. The available results are visualised via maps with colourbars.Öğe Spreading in scale-free computer networks with improved clustering(World Scientific Publ Co Pte Ltd, 2018) Turker, Ilker; Albayrak, ZaferIn this study, we investigated data spreading in computer networks with scale-free topology under various levels of improved clustering. Starting from a pure Barabasi-Albert (BA) network topology, we applied a Poisson-based rewiring procedure with increasing rewiring probability, which promotes local connections. We then performed wired computer network simulations in NS2 simulator for these topologies. We found that for pure BA network, data transfer (throughput) is maximum, where time required for establishing routing scheme, end-to-end delays in data transmission and number of nodes acting in data transfer are at their minimum levels. Improving clustering increases these parameters those are at their minima. A noteworthy finding of this study is that, for moderate levels of clustering, total throughput remains close to its maximum yielding stable transfer rates, although number of infected nodes and end-to-end delay increase. This indicates that clustering promotes spreading phenomena in networks, although it increases average separation. As a result, clustering property emerges as a catalyzer in data spreading with minimal effects on the total amount of transmission.Öğe TOOL WEAR PREDICTION BY DEEP LEARNING FROM AUGMENTABLE VISIBILITY GRAPH REPRESENTATION OF TIME SERIES DATA(Technical Univ Cluj-Napoca, Fac Machine Building Dept Systems Eng, 2023) Turker, Ilker; Tan, Serhat Orkun; Kutluana, GokhanTool wear prediction has a crucial role for improving manufacturing quality and reliability due to optimizing tool replacement schedules, reducing downtime, and improving overall production efficiency. Deep learning models, having the ability to analyze large and complex datasets, can extract relevant information, and make accurate predictions about the condition of cutting tools. We propose a smart detection methodology based on converting the available sensory data collected from a CNC milling machine into a visibility graph representation. Due to the high dimensionality of the data with 44 attributes related to machining, a multilayer visibility graph representation is achieved after this conversion procedure, resulting in a 44-layered 128x128 adjacency matrix formation. A novel data augmentation technique specifically applicable to graph representation is also employed to increase the data size originally composed of 18 experiments into 360, each one represented as a multilayer graph. Augmented graph representations are further input to a custom CNN deep learning architecture with a split of 70% train, 10% validation and 20% test instances. Results indicate that Augmented Graph-induced classification of CNC mill tool with custom CNN model (GA-CNN) yields full accuracy for detecting whether the tool is worn or not.Öğe Uncovering the differences in linguistic network dynamics of book and social media texts(Springer International Publishing Ag, 2016) Turker, Ilker; Sehirli, Eftal; Demiral, EmrullahComplex network studies span a large variety of applications including linguistic networks. To investigate the differences in book and social media texts in terms of linguistic typology, we constructed both sequential and sentence collocation networks of book, Facebook and Twitter texts with undirected and weighted edges. The comparisons are performed using the basic parameters like average degree, modularity, average clustering coefficient, average path length, diameter, average link weight etc. We also presented the distribution graphs for node degrees, edge weights and maximum degree differences of the pairing nodes. The degree difference occurrences are furtherly detailed with the grayscale percentile plots with respect to the edge weights. We linked the network analysis with linguistic aspects like word and sentence length distributions. We concluded that linguistic typology demonstrates a formal usage in book that slightly deviates to informal in Twitter. Facebook interpolates between these media by the means of network parameters, while the informality of Twitter is mostly influenced by the character limitations.
