Zeynep Turgut - Academia.edu (original) (raw)
Papers by Zeynep Turgut
2023 14th International Conference on Network of the Future (NoF)
Future Generation Computer Systems
Journal of Ambient Intelligence and Smart Environments, Mar 27, 2023
Network Modeling Analysis in Health Informatics and Bioinformatics
2022 9th International Conference on Future Internet of Things and Cloud (FiCloud)
Journal of Ambient Intelligence and Smart Environments, Mar 14, 2023
Kocaeli Universitesi, Fen Bilimleri Enstitusu, Jun 25, 2009
Lecture Notes in Electrical Engineering, 2018
In today’s world, where all records are carried into an electronic environment, cyber security re... more In today’s world, where all records are carried into an electronic environment, cyber security represents a very broad scope, with the primary objective of preventing the loss of financial and/or emotional loss of people, institutions, organizations through the security of data in the digital environment. Today, the most common cyber security threat is phishing attacks. With the phishing attack, the attacker aims to capture the data which are very important for the individuals like identification number, social security number, bank account information, and so on. In this study, using deep learning, it was checked whether the web sites are real or not by using neural networks and support vector machine, decision tree and stacked autoencoders as classification methods. As a result of the study, 86% success rate was reached by using stacked autoencoders which are a part of deep learning techniques.
Lecture Notes in Electrical Engineering, 2018
In this study, the indoor localization was performed on indoor networks. WiFi technology is locat... more In this study, the indoor localization was performed on indoor networks. WiFi technology is located in almost every building. For this reason, WiFi technology has been selected to perform positioning, and RSSI values from WiFi technology access points have been examined. For this purpose, RFKON_MB_WIFI dataset in RFKON database which is a sample database is used and data of 18480 signal strength are analyzed. The Fingerprinting method of Scene Analysis methods is used to perform the localization process. As a first step, the signal strengths in the data set are normalized by preprocessing. In the second step, positioning was performed using SVM, PCA, LDA, KNN, N3, BNN, Naive Bayes Classification, and Deep Learning methods. When the results obtained are compared, the most successful result is obtained from deep learning that is known to have a high accuracy on big data with an accuracy of 95.95%.
As the accuracy rate of artificial intelligence based applications increased, they have started t... more As the accuracy rate of artificial intelligence based applications increased, they have started to be used in different areas. Artifical Neural Networks (ANN) can be very successful for extracting meaningful data from features by processing complex data. Well-trained models can solve difficult problems with high a high accuracy rate. In this study, 2 different ANN models have been developed to detect malicious users who want to access high-security servers. These models are tested from simple to complex: Shallow Neural Network (SNN), Deep Neural Network (DNN), and Auto Encoder are used to reduce features. All models are trained with CICIDS2017 dataset. Server connection requests are classified as normal or malicious (Brute Force, Web Attack, In ltration, Botnet or DDoS) with 98.45% accuracy rate.
In this study, the indoor localization was performed on indoor networks. WiFi technology is locat... more In this study, the indoor localization was performed on indoor networks. WiFi technology is located in almost every building. For this reason, WiFi technology has been selected to perform positioning, and RSSI values from WiFi technology access points have been examined. For this purpose, RFKON_MB_WIFI dataset in RFKON database which is a sample database is used and data of 18480 signal strength are analyzed. The Fingerprinting method of Scene Analysis methods is used to perform the localization process. As a first step, the signal strengths in the data set are normalized by preprocessing. In the second step, positioning was performed using SVM, PCA, LDA, KNN, N3, BNN, Naive Bayes Classification, and Deep Learning methods. When the results obtained are compared, the most successful result is obtained from deep learning that is known to have a high accuracy on big data with an accuracy of 95.95%.
In this study, Earth’s magnetic field signals have been investigated to determine mobile user’s l... more In this study, Earth’s magnetic field signals have been investigated to determine mobile user’s location. In theory, Earth’s magnetic field does not change during the day at a certain point. But, the various noise effects that are exposed during the measurement causes deviations in the measured signal. In this study; Kalman Filter, LOESS, Savitzky-Golay filters are adapted with two different approaches to purge Earth’s magnetic field values from noise. KNearest Neighbour and Random Forest models have been trained with filtered signals and the locations of the mobile user are determined. Relevant systems have been tested by using RFKONDB which is existed in literature. The purpose of this study is to measure how these filters should be adapted to an Earth’s magnetic field based indoor localization systems. Digital sensors, which are integrated mobile devices, can use different measurement techniques. In a heterogeneous environment, noise reduction filters can show a different effect....
Identification systems have become biometric-based, especially with the increase in the performan... more Identification systems have become biometric-based, especially with the increase in the performance rates of machine learning methods. Biometric identification systems offer a high level of security by using reliable, difficult-to-change parameters. In this thesis, a biometric identification system is proposed using dorsal hand vein patterns. The relevant system has been tested on the sample dataset in the literature. The number of data were increased by adding noisy data to the data set used. The classification was made on the preprocessed images using SVM, ANN, LDA + KNN, and CNN methods. It has been determined that the highest identification accuracy is achieved when CNN is used, and CNN method provides higher performance compared to other methods. With the proposed identification system, after multiple runs, an average accuracy of 99.64% is achieved with the CNN machine learning method.
2018 International Congress on Big Data, Deep Learning and Fighting Cyber Terrorism (IBIGDELFT)
Transactions on Emerging Telecommunications Technologies
Procedia Computer Science, 2016
Abstract “Knowing the Location” and “Determining the Location” are the essential requirements of ... more Abstract “Knowing the Location” and “Determining the Location” are the essential requirements of constructing a smart building. GPS technology, which is often used for the purpose of positioning, cannot be used efficiently while performing position detection indoors because of the losses occurring in the signal propagation. At the same time, the object of sensors, Bluetooth, IrDA and RFID devices which are commonly used in a building, and bandwidth constraints makes such location detection more difficult because of possible energy consumption and limited memory capacity. Since WiFi technology which will probably be used in nearly all Smart Buildings, Indoor localization algorithms have been surveyed and requirements which are essential for obtaining Internet of Things (IoT) technology have been researched, and a context-based approach for a Smart Building which has got IoT structure is proposed.
2023 14th International Conference on Network of the Future (NoF)
Future Generation Computer Systems
Journal of Ambient Intelligence and Smart Environments, Mar 27, 2023
Network Modeling Analysis in Health Informatics and Bioinformatics
2022 9th International Conference on Future Internet of Things and Cloud (FiCloud)
Journal of Ambient Intelligence and Smart Environments, Mar 14, 2023
Kocaeli Universitesi, Fen Bilimleri Enstitusu, Jun 25, 2009
Lecture Notes in Electrical Engineering, 2018
In today’s world, where all records are carried into an electronic environment, cyber security re... more In today’s world, where all records are carried into an electronic environment, cyber security represents a very broad scope, with the primary objective of preventing the loss of financial and/or emotional loss of people, institutions, organizations through the security of data in the digital environment. Today, the most common cyber security threat is phishing attacks. With the phishing attack, the attacker aims to capture the data which are very important for the individuals like identification number, social security number, bank account information, and so on. In this study, using deep learning, it was checked whether the web sites are real or not by using neural networks and support vector machine, decision tree and stacked autoencoders as classification methods. As a result of the study, 86% success rate was reached by using stacked autoencoders which are a part of deep learning techniques.
Lecture Notes in Electrical Engineering, 2018
In this study, the indoor localization was performed on indoor networks. WiFi technology is locat... more In this study, the indoor localization was performed on indoor networks. WiFi technology is located in almost every building. For this reason, WiFi technology has been selected to perform positioning, and RSSI values from WiFi technology access points have been examined. For this purpose, RFKON_MB_WIFI dataset in RFKON database which is a sample database is used and data of 18480 signal strength are analyzed. The Fingerprinting method of Scene Analysis methods is used to perform the localization process. As a first step, the signal strengths in the data set are normalized by preprocessing. In the second step, positioning was performed using SVM, PCA, LDA, KNN, N3, BNN, Naive Bayes Classification, and Deep Learning methods. When the results obtained are compared, the most successful result is obtained from deep learning that is known to have a high accuracy on big data with an accuracy of 95.95%.
As the accuracy rate of artificial intelligence based applications increased, they have started t... more As the accuracy rate of artificial intelligence based applications increased, they have started to be used in different areas. Artifical Neural Networks (ANN) can be very successful for extracting meaningful data from features by processing complex data. Well-trained models can solve difficult problems with high a high accuracy rate. In this study, 2 different ANN models have been developed to detect malicious users who want to access high-security servers. These models are tested from simple to complex: Shallow Neural Network (SNN), Deep Neural Network (DNN), and Auto Encoder are used to reduce features. All models are trained with CICIDS2017 dataset. Server connection requests are classified as normal or malicious (Brute Force, Web Attack, In ltration, Botnet or DDoS) with 98.45% accuracy rate.
In this study, the indoor localization was performed on indoor networks. WiFi technology is locat... more In this study, the indoor localization was performed on indoor networks. WiFi technology is located in almost every building. For this reason, WiFi technology has been selected to perform positioning, and RSSI values from WiFi technology access points have been examined. For this purpose, RFKON_MB_WIFI dataset in RFKON database which is a sample database is used and data of 18480 signal strength are analyzed. The Fingerprinting method of Scene Analysis methods is used to perform the localization process. As a first step, the signal strengths in the data set are normalized by preprocessing. In the second step, positioning was performed using SVM, PCA, LDA, KNN, N3, BNN, Naive Bayes Classification, and Deep Learning methods. When the results obtained are compared, the most successful result is obtained from deep learning that is known to have a high accuracy on big data with an accuracy of 95.95%.
In this study, Earth’s magnetic field signals have been investigated to determine mobile user’s l... more In this study, Earth’s magnetic field signals have been investigated to determine mobile user’s location. In theory, Earth’s magnetic field does not change during the day at a certain point. But, the various noise effects that are exposed during the measurement causes deviations in the measured signal. In this study; Kalman Filter, LOESS, Savitzky-Golay filters are adapted with two different approaches to purge Earth’s magnetic field values from noise. KNearest Neighbour and Random Forest models have been trained with filtered signals and the locations of the mobile user are determined. Relevant systems have been tested by using RFKONDB which is existed in literature. The purpose of this study is to measure how these filters should be adapted to an Earth’s magnetic field based indoor localization systems. Digital sensors, which are integrated mobile devices, can use different measurement techniques. In a heterogeneous environment, noise reduction filters can show a different effect....
Identification systems have become biometric-based, especially with the increase in the performan... more Identification systems have become biometric-based, especially with the increase in the performance rates of machine learning methods. Biometric identification systems offer a high level of security by using reliable, difficult-to-change parameters. In this thesis, a biometric identification system is proposed using dorsal hand vein patterns. The relevant system has been tested on the sample dataset in the literature. The number of data were increased by adding noisy data to the data set used. The classification was made on the preprocessed images using SVM, ANN, LDA + KNN, and CNN methods. It has been determined that the highest identification accuracy is achieved when CNN is used, and CNN method provides higher performance compared to other methods. With the proposed identification system, after multiple runs, an average accuracy of 99.64% is achieved with the CNN machine learning method.
2018 International Congress on Big Data, Deep Learning and Fighting Cyber Terrorism (IBIGDELFT)
Transactions on Emerging Telecommunications Technologies
Procedia Computer Science, 2016
Abstract “Knowing the Location” and “Determining the Location” are the essential requirements of ... more Abstract “Knowing the Location” and “Determining the Location” are the essential requirements of constructing a smart building. GPS technology, which is often used for the purpose of positioning, cannot be used efficiently while performing position detection indoors because of the losses occurring in the signal propagation. At the same time, the object of sensors, Bluetooth, IrDA and RFID devices which are commonly used in a building, and bandwidth constraints makes such location detection more difficult because of possible energy consumption and limited memory capacity. Since WiFi technology which will probably be used in nearly all Smart Buildings, Indoor localization algorithms have been surveyed and requirements which are essential for obtaining Internet of Things (IoT) technology have been researched, and a context-based approach for a Smart Building which has got IoT structure is proposed.