Browsing by Author "Akbas, Ayhan"

Now showing 1 - 9 of 9

A Comparative Analysis of Passenger Flow Forecasting in Trams Using Machine Learning Algorithms
(Bitlis Eren Üniversitesi, 2024) Adanur Dedeturk, Beyhan; Dedeturk, Bilgi Kağan; Akbas, Ayhan; 0000-0003-4983-2417; AGÜ, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü; Adanur Dedeturk, Beyhan
Forecasting tram passenger flow is an important part of the intelligent transportation system since it helps with resource allocation, network design, and frequency setting. Due to varying destinations and departure times, it is difficult to notice large fluctuations, non-linearity, and periodicity of tram passenger flows. In this paper, the first-order difference technique is used to eliminate seasonal structure from the time series data and the performance of different machine learning algorithms on passenger flow forecasting in trams is evaluated. Furthermore, the impact of the Covid-19 pandemic on forecasting success is examined. For this purpose, the tram data of Kayseri Transportation Inc. for the years 2018-2021 are used. Different estimation models including Linear Regression, Support Vector Regression, Random Forest, Artificial Neural Network, Convolutional Neural Network, and LongTerm Short Memory are applied and the time series forecasting performances of the models are evaluated with MAPE and R2 metrics.
COMPARATIVE PERFORMANCE ANALYSIS OF ARIMA, PROPHET AND HOLT-WINTERS FORECASTING METHODS ON EUROPEAN COVID-19 DATA
(Kerim ÇETİNKAYA, 2022) Ersöz, Nur Şebnem; Güner Şahan, Pınar; Akbas, Ayhan; Bakır Güngör, Burcu; 0000-0003-3343-9936; 0000-0001-5979-0375; 0000-0002-6425-104X; 0000-0002-2272-6270; AGÜ, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü; Ersöz, Nur Şebnem; Güner Şahan, Pınar; Akbas, Ayhan; Bakır Güngör, Burcu
COVID-19 is the most common infectious disease of the last few years and has caused an outbreak all around the world. The mortality rate, which was earlier in the hundreds, increased to thousands and then to millions. Since January 2020, several scientists have attempted to understand and predict the spread of COVID-19 so that governments may make sufficient arrangements in hospitals and reduce the number of deaths. This research article presents a comparative performance analysis of ARIMA, Prophet and HoltWinters Exponential Smoothing forecasting methods to make predictions for the COVID-19 disease epidemiology in Europe. The dataset has been collected from the World Health Organization (WHO) and includes the COVID-19 case data of European countries, which is categorized by WHO between the years of 2020 and 2022. The results indicate that Holt-Winters Exponential Smoothing method (RMSE: 0.2080, MAE: 0.1747) outperforms ARIMA and Prophet forecasting methods.
Document Classification with Contextually Enriched Word Embeddings
(Bajece (İstanbul Teknik Ünv), 2024) Mahmood, Raad Saadi; Bakal, Gokhan; Akbas, Ayhan; 0000-0003-2897-3894; AGÜ, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü; Bakal, Mehmet Gokhan
The text classification task has a wide range of application domains for distinct purposes, such as the classification of articles, social media posts, and sentiments. As a natural language processing application, machine learning and deep learning techniques are intensively utilized in solving such challenges. One common approach is employing the discriminative word features comprising Bag-of-Words and n-grams to conduct text classification experiments. The other powerful approach is exploiting neural network-based (specifically deep learning models) through either sentence, word, or character levels. In this study, we proposed a novel approach to classify documents with contextually enriched word embeddings powered by the neighbor words accessible through the trigram word series. In the experiments, a well-known web of science dataset is exploited to demonstrate the novelty of the models. Consequently, we built various models constructed with and without the proposed approach to monitor the models' performances. The experimental models showed that the proposed neighborhood-based word embedding enrichment has decent potential to use in further studies.
Human identification using palm print images based on deep learning methods and gray wolf optimization algorithm
(SPRINGER, 2024) Alshakree, Firas; Akbas, Ayhan; Rahebi, Javad; 0000-0002-6425-104X; AGÜ, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü; Akbas, Ayhan
Palm print identification is a biometric technique that relies on the distinctive characteristics of a person’s palm print to distinguish and authenticate their identity. The unique pattern of ridges, lines, and other features present on the palm allows for the identification of an individual. The ridges and lines on the palm are formed during embryonic development and remain relatively unchanged throughout a person’s lifetime, making palm prints an ideal candidate for biometric identification. Using deep learning networks, such as GoogLeNet, SqueezeNet, and AlexNet combined with gray wolf optimization, we achieved to extract and analyze the unique features of a person’s palm print to create a digital representation that can be used for identification purposes with a high degree of accuracy. To this end, two well-known datasets, the Hong Kong Polytechnic University dataset and the Tongji Contactless dataset, were used for testing and evaluation. The recognition rate of the proposed method was compared with other existing methods such as principal component analysis, including local binary pattern and Laplacian of Gaussian-Gabor transform. The results demonstrate that the proposed method outperforms other methods with a recognition rate of 96.72%. These findings show that the combination of deep learning and gray wolf optimization can effectively improve the accuracy of human identification using palm print images.
Machine learning approaches for underwater sensor network parameter prediction
(ELSEVIER, 2023) Uyan, Osman Gokhan; Akbas, Ayhan; Gungor, Vehbi Cagri; 0000-0003-3922-1647; 0000-0002-6425-104X; 0000-0003-0803-8372; AGÜ, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü; Uyan, Osman Gokhan; Akbas, Ayhan; Gungor, Vehbi Cagri
Underwater Acoustic Sensor Networks (UASNs) have recently attracted scientists due to its wide range of real -world applications. However, there are design challenges in UASNs, such as limited network lifetime and low communication reliability provoked by the constrained battery supply of sensors and harsh channel conditions in the underwater environments. To meet communication reliability requirements, packet-duplication and multi -path routing algorithms have been recommended in the literature. Furthermore, underwater sensors may convey sensitive data, which must be masked to avoid eavesdropping attempts. To improve network security, cryptographic encryption is the most widely used method. Nevertheless, data encryption needs computations to cipher the data, which consumes extra energy, resulting in a cutback in the life span of the network. To address these challenges, an optimization model has been proposed to evaluate the impacts of multi-path routing, packet duplication, encryption, and data fragmentation on the lifetime of the UASNs. However, the solution time of the proposed optimization model is quite high, and sometimes it cannot come up with feasible solutions. To this end, in this study, different regression and neural network methods have been proposed to predict network param-eters and energy consumptions of underwater nodes as supplementary methods to optimization models. Per-formance evaluations show that the proposed methods yield remarkably accurate predictions and can be used for energy consumption prediction in UASNs.
Machine Learning based Early Prediction of Type 2 Diabetes: A New Hybrid Feature Selection Approach using Correlation Matrix with Heatmap and SFS
(İstanbul Teknik Üniversitesi, 2022) Akbas, Ayhan; Buyrukoğlu, Selim; 0000-0002-6425-104X; AGÜ, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü; Akbas, Ayhan
A new hybrid machine learning method for the prediction of type 2 diabetes is introduced and explained in detail. Also outcomes are compared with the similar researches. Early prediction of diabetes is crucial to take necessary measures (i.e. changing eating habits, patient weight control etc.), to defer the emergence of diabetes and to reduce the death rate to some extent and ease medical care professionals’ decision making in preventing and managing diabetes mellitus.The purpose of this study is the creation of a new hybrid feature selection approach combination of Correlation Matrix with Heatmap and Sequential forward selection (SFS) to reveal the most effective features in the detection of diabetes. A diabetes data set with 520 instances and seven features were studied with the application of the proposed hybrid feature selection approach. The evaluation of the selected optimal features was measured by applying Support Vector Machines(SVM), Random Forest(RF), and Artificial Neural Networks(ANN) classifiers. Five evaluation metrics, namely, Accuracy, F-measure, Precision, Recall, and AUC showed the best performance with ANN (99.1%), F-measure (99.1%), Precision (99.3%), Recall (99.1%), and AUC (99.8%). Our proposed hybrid feature selection model provided a more promising performance with ANN compared to other machine learning algorithms.
Machine Learning based Network Intrusion Detection with Hybrid Frequent Item Set Mining
(GAZİ ÜNİVERSİTESİ, 2024) Fırat, Murat; Bakal, Gokhan; Akbas, Ayhan; 0000-0003-2897-3894; AGÜ, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü; Bakal, Gokhan
With the development and expansion of computer networks day by day and the diversity of software developed, the damage that possible attacks can cause is increasing beyond the predictions. Intrusion Detection Systems (STS/IDS) are one of the practical defense tools against these potential attacks that are constantly growing and diversifying. Thus, one of the emerging methods among researchers is to train these systems with various artificial intelligence methods to detect subsequent attacks in real time and take the necessary precautions. However, the ultimate goal is to propose a hybrid feature selection approach to improve the classification performance. The raw dataset originally enclosed 85 descriptor features (attributes) for classification. These attributes are extracted using CICFlowMeter from a PCAP file where network traffic is recorded for data curation. In this study, classical feature selection methods and frequent item set mining approaches were employed in feature selection for constructing a hybrid model. We aimed to examine the effect of the proposed hybrid feature selection approach on the classification task for the network traffic data containing ordinary and attack records. The outcomes demonstrate that the proposed method gained nearly 3% improvement when applied with the Logistic Regression algorithm on classifying more than 225,000 records.
A reliable and secure multi-path routing strategy for underwater acoustic sensor networks
(ELSEVIER, 2022) Uyan, Osman Gokhan; Akbas, Ayhan; Gungor, Vehbi Cagri; 0000-0003-3922-1647; 0000-0002-6425-104X; 0000-0003-0803-8372; AGÜ, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü; Uyan, Osman Gokhan; Akbas, Ayhan; Gungor, Vehbi Cagri
Underwater Acoustic Sensor Networks (UASNs) have nowadays become an attractive topic in scientific studies and commercial applications. An important challenge in UASN’s design is the limited network lifetime and low reliability caused by the limited battery energy of sensor nodes and harsh channel conditions in underwater environments. In addition, sensor nodes may generate sensitive data, which needs to be concealed. To this end, cryptographic encryption is a commonly used method to cipher a data before transmission to maintain security. However, encryption methods require additional computation and extra energy, which causes a decrease in the network lifetime. To this end, transmitting fragmented data through multiple paths can be used as a security countermeasure, in conjunction with encryption against silent listening attacks. To address these challenges, in this study, an optimization framework has been developed to analyze the effects of multi-path routing, packet duplication, encryption and data fragmentation on network lifetime. In addition to an optimal solution, Simulated Annealing, Golden Section Search and Genetic Algorithm-based heuristic methods have been developed. Performance results show that the proposed approach jointly solves the problem of UASN lifetime maximization, while providing network reliability and security
Stacking Ensemble Learning-Based Wireless Sensor Network Deployment Parameter Estimation
(Institute for Ionics, 2023) Akbas, Ayhan; Buyrukoglu, Selim; 0000-0002-6425-104X; AGÜ, Mühendislik Fakültesi, Bilgisayar Mühendisliği Bölümü; Akbas, Ayhan
In wireless sensor network projects, it is generally desired to cover the area to be monitored at a given cost and to achieve the maximum useful network lifetime. In the deployment of the wireless sensors, it is necessary to know in advance how many sensor nodes will be required, how much the distance between the nodes should be, etc., or what the transmit power level should be, etc. depending on the channel parameters of the area. This necessitates accurate calculation of variables such as maximum network lifetime, communication channel parameters, number of nodes to be used, and distance between nodes. As numbers reach to the order of hundreds, calculation tends to a NP hard problem to solve. At this point, we employed both single-based and stacked ensemble-based machine learning models to speed up the parameter estimations with highly accurate outcomes. Adaboost was superior over other models (Elastic Net, SVR) in single-based models. Stacked ensemble models achieved best results for the WSN parameter prediction compared to single-based models.