Araştırma Çıktıları | WoS | Scopus | TR-Dizin | PubMed
Permanent URI for this communityhttps://hdl.handle.net/20.500.14719/1741
Browse
26 results
Search Results
Publication Metadata only Delivering Machine Learning Applications via Cloud Platforms: An Experience Report(Institute of Electrical and Electronics Engineers Inc., 2020) Sener, Yigit; Yetim, Hasan Fahri; Baǧriyanik, Selami; Akgun, B.T.; Ayav, T.; Bilgen, S.; Kardas, G.; Sener, Yigit, M.Sc. Program in Big Data Analytics and Management, Bahçeşehir Üniversitesi, Istanbul, Turkey; Yetim, Hasan Fahri, M.Sc. Program in Big Data Analytics and Management, Bahçeşehir Üniversitesi, Istanbul, Turkey; Baǧriyanik, Selami, Turkcell Teknoloji, Digital Services and Solutions Technology Department, Istanbul, TurkeyCloud technologies enable developers and organizations to focus on their product, without having to consider issues such as local server capacity, infrastructure modifications, data security, licensing or human capital. This paper attempts to explain a case in which a Machine Learning application is deployed via Amazon Web Services (AWS) tools. In doing so, it demonstrates the reasoning behind choosing a cloud-based environment instead of on-premise sources, by putting forward the advantages of the former. On the other hand, it should be noted that the application in this experience is generated with a hybrid approach: It is developed using on-premise infrastructure and then moved to the Cloud environment for the deployment phase only. In this regard, it can be read as a PaaS experience. This study is considered to be a beneficial guide for entrepreneurs and start-ups on a budget who aim at launching their products in a swift and scalable manner. © 2020 Elsevier B.V., All rights reserved.Publication Metadata only Automating Customer Claim Registration by Text Mining(Institute of Electrical and Electronics Engineers Inc., 2020) Beyranvand, Peyman; Aytekin, Tevfik; Beyranvand, Peyman, Research and Development Mayen CRM, Istanbul, Turkey; Aytekin, Tevfik, Department of Computer Engineering, Bahçeşehir Üniversitesi, Istanbul, TurkeyIn this paper, we present the use of text mining and machine learning in call centers to increase the efficiency of registering customer claims and improving customer satisfaction. Our proposed method makes the process of claim registration faster and more accurate compared to experienced call center agents. Use of text mining and machine learning techniques will increase the customer satisfaction and endows the call center staff with better ways to help the customer. © 2020 Elsevier B.V., All rights reserved.Publication Metadata only Tor Network Detection by Using Machine Learning and Artificial Neural Network(Institute of Electrical and Electronics Engineers Inc., 2021) Soykan, Murat; Boluk, Pinar Sarisaray; Soykan, Murat, Computer Engineering, Bahçeşehir Üniversitesi, Istanbul, Turkey; Boluk, Pinar Sarisaray, Department of Software Engineering, Bahçeşehir Üniversitesi, Istanbul, TurkeyThe Internet is a virtual world where everyone can express themselves as much as they wish and perform the operations they want. In this virtual world, some users want to experience the internet without giving their identity for certain reasons. The concept of an anonymous network has emerged so that they can use the internet without revealing their identity. The Tor project is a product that provides anonymous communication on the Internet without revealing users' identities. In this project, we aimed to determine whether network traffic is the TOR network by using machine learning and artificial neural networks. With the dataset we have, we first performed data analysis and gained more information about the data set. Categorical values were assigned to numerical values to learn the dataset. After converting the categorical data to numerical data, normalization is applied to the data set and all features are taken between -1 and 1. It was estimated whether the future traffic was TOR by learning the past data by using K Nearest Neighbor, Naive Bayes Classifiers, Random Decision Forest, Logistic Regression, Support Vector Machine, one of the machine learning classification algorithms. In addition, artificial neural networks were used. After each algorithm, confusion matrix, precision, recall, and F1-score values, which are among the model evaluation tools, were calculated, and compared which model performed better for our dataset. © 2022 Elsevier B.V., All rights reserved.Publication Metadata only A Review of Spam Detection in Social Media(Institute of Electrical and Electronics Engineers Inc., 2021) Yurtseven, Ilke; Baǧriyanik, Selami; Ayvaz, Serkan; Yurtseven, Ilke, Bahçeşehir Üniversitesi, Istanbul, Turkey; Baǧriyanik, Selami, Department of Software Engineering, Nişantaşı Üniversitesi, Istanbul, Turkey; Ayvaz, Serkan, Department of Computer Engineering, Yıldız Teknik Üniversitesi, Istanbul, TurkeyWith significant usage of social media to socialize in virtual environments, bad actors are now able to use these platforms to spread their maUcious activities such as hate speech, spam, and even phishing to very large crowds. Especially, Twitter is suitable for these types of activities because it is one of the most common social media platforms for microblogging with millions of active users. Moreover, since the end of 2019, Covid-19 has changed the lives of individuals in many ways. While it increased social media usage due to free time, the number of cyber-attacks soared too. To prevent these activities, detection is a very crucial phase. Thus, the main goal of this study is to review the state-of-art in the detection of malicious content and the contribution of AI algorithms for detecting spam and scams effectively in social media. © 2022 Elsevier B.V., All rights reserved.Publication Metadata only A Touchless Control Interface for Low-Cost ROVs(Institute of Electrical and Electronics Engineers Inc., 2021) Kapicioglu, Kagan; Getmez, Enis; Akbulut, Batuhan Ekin; Akgul, Arda; Ucar, Burak; Kanlikilic, Berke; Koc, Mehmet; Gür, M. Berke; Kapicioglu, Kagan, Mechatronics Engineering, Bahçeşehir Üniversitesi, Istanbul, Turkey; Getmez, Enis, Software Development, Caretta Yazilim Ltd. Şti., Istanbul, Turkey; Akbulut, Batuhan Ekin, Mechatronics Engineering, Bahçeşehir Üniversitesi, Istanbul, Turkey; Akgul, Arda, Mechatronics Engineering, Bahçeşehir Üniversitesi, Istanbul, Turkey; Ucar, Burak, Mechatronics Engineering, Bahçeşehir Üniversitesi, Istanbul, Turkey; Kanlikilic, Berke, Mechatronics Engineering, Bahçeşehir Üniversitesi, Istanbul, Turkey; Koc, Mehmet, Department of Electrical and Electronic Engineering, Bahçeşehir Üniversitesi, Istanbul, Turkey; Gür, M. Berke, Mechatronics Engineering, Bahçeşehir Üniversitesi, Istanbul, TurkeyIn this paper, a fully touchless control interface for low cost remotely operated vehicles (ROVs) is presented. This interface aims to decrease training time, reduce workload, and ensure the operation ergonomics of ROV operators. Fully touchless control interface is achieved by a machine learning (ML) algorithm for ROV operator's face and orientation recognition, and controling the angle of an ROV-based camera. Furthermore, a Leap Motion sensor captures hand gestures and movements, thereby allowing the ROV operator to execute maneuvers and perform other functions (e.g., gripper or lighting control) based on pre-determined hand gestures. Fusion of face and hand gestures allows the operator to control ROV in a fully touchless way. The proposed system is tested in a realistic underwater simulation environment designed specifically for typical tasks that are present in student competitions. Trials with inexperienced operators show that the touchless interface can cut training times, speed up operations, reduce workload, and can provide the operator with a more natural feeling of command and control as well as better ergonomy. © 2022 Elsevier B.V., All rights reserved.Publication Metadata only The analysis of feature selection with machine learning for indoor positioning(Institute of Electrical and Electronics Engineers Inc., 2021) Aydin, Hurkan M.; Ali, Muhammad Ammar; Gelal, Ece; Aydin, Hurkan M., Department of Electrical and Electronic Engineering, Bahçeşehir Üniversitesi, Istanbul, Turkey; Ali, Muhammad Ammar, Computer Engineering, Bahçeşehir Üniversitesi, Istanbul, Turkey; Gelal, Ece, Computer Engineering, Bahçeşehir Üniversitesi, Istanbul, TurkeyIndoor positioning is useful in various venues including warehouses, convention centers, malls, airports, nursing homes. In these scenarios, reducing the complexity of location estimation both improves responsiveness and helps to elongate battery life of the mobile device. In this work, we carry out a detailed analysis of the impact of Principal Component Analysis (PCA) on the computational complexity and accuracy with different machine learning algorithms on a large data set containing 520 APs. We compare the algorithms' training and testing times, as well as their accuracies in the presence and absence of PCA. Our results show that (i) PCA significantly reduces both the training and testing times for classification and regression using k-nearest neighbor (kNN) and support vector machine (SVM) algorithms while preserving if not improving accuracy, (ii) PCA slightly improves the training/testing times for regression using multi-layer perceptron (MLP), (iii) random forest (RF) does not perform well with PCA. © 2021 Elsevier B.V., All rights reserved.Publication Metadata only Credit risk analysis using machine-learning algorithms, Makine öǧrenmesi algoritmalarini kullanarak kredi riski analizi(Institute of Electrical and Electronics Engineers Inc., 2021) Alagoz, Gokhan; Çanakoǧlu, Ethem; Alagoz, Gokhan, Kuveyt Türk Katilim Bankasi, Kocaeli, Turkey; Çanakoǧlu, Ethem, Bahçeşehir Üniversitesi, Istanbul, TurkeyCredit risk and default risk are very important concepts for all banks and financial institutions globally. As credit risk measurement and modeling requires working with large samples, it was preferred to use machine learning, one of the modern analysis techniques. In the study, Logistic Regression, Random Forest and Artificial Neural Networks, which are quite common among machine learning algorithms and have taken their place in research, were used for data analysis and modeling studies. 66,078 loan samples and 11 variables belonging to these samples have been used in the modeling, and research has been carried out on the estimation of the results of successful (0) or unsuccessful (1) credits using machine learning algorithms according to the effect of these variables. These records were taken anonymously from the test environment of a financial institution and the necessary data engineering and modeling studies were carried out using the Python programming language. As a result of the study, it was seen that the model established with Logistic Regression produces better results and is a more suitable method for the research subject. © 2021 Elsevier B.V., All rights reserved.Publication Metadata only Machine Vision-Based Expert System for Automated Skin Cancer Detection(Springer Science and Business Media Deutschland GmbH, 2022) Junayed, Masum Shah; Jeny, Afsana Ahsan; Rada, Lavdie; Islam, Md Baharul; Brito-Loeza, C.; Martin-Gonzalez, A.; Castañeda-Zeman, V.; Safi, A.; Junayed, Masum Shah, Department of Computer Engineering, Bahçeşehir Üniversitesi, Istanbul, Turkey, Department of CSE, Daffodil International University, Dhaka, Bangladesh; Jeny, Afsana Ahsan, Department of Computer Engineering, Bahçeşehir Üniversitesi, Istanbul, Turkey; Rada, Lavdie, Department of Computer Engineering, Bahçeşehir Üniversitesi, Istanbul, Turkey; Islam, Md Baharul, Department of Computer Engineering, Bahçeşehir Üniversitesi, Istanbul, Turkey, College of Data Science and Engineering, American University of Malta, Cospicua, MaltaSkin cancer is the most frequently occurring kind of cancer, accounting for about one-third of all cases. Automatic early detection without expert intervention for a visual inspection would be of great help for society. The image processing and machine learning methods have significantly contributed to medical and biomedical research, resulting in fast and exact inspection in different problems. One of such problems is accurate cancer detection and classification. In this study, we introduce an expert system based on image processing and machine learning for skin cancer detection and classification. The proposed approach consists of three significant steps: pre-processing, feature extraction, and classification. The pre-processing step uses the grayscale conversion, Gaussian filter, segmentation, and morphological operation to represent skin lesion images better. We employ two feature extractors, i.e., the ABCD scoring method (asymmetry, border, color, diameter) and gray level co-occurrence matrix (GLCM), to extract cancer-affected areas. Finally, five different machine learning classifiers such as logistic regression (LR), decision tree (DT), k-nearest neighbors (KNN), support vector machine (SVM), and random forest (RF) used to detect and classify skin cancer. Experimental results show that random forest exceeds all other classifiers achieving an accuracy of 97.62% and 0.97 Area Under Curve (AUC), which is state-of-the-art on the experimented open-source dataset PH2. © 2022 Elsevier B.V., All rights reserved.Publication Metadata only Toward Automatic Water Pollution Analysis: A Machine Learning Approach for Water-Quality Monitoring Through Pattern Classification of Water Crystallization(Springer Science and Business Media Deutschland GmbH, 2022) Rada, Lavdie; Tanrıverdi, Yusuf Baran; Kara, Ömer Ekmel; Hemond, Elizabeth M.; Tezel, Ulaş; Brito-Loeza, C.; Martin-Gonzalez, A.; Castañeda-Zeman, V.; Safi, A.; Rada, Lavdie, Bahçeşehir Üniversitesi, Istanbul, Turkey; Tanrıverdi, Yusuf Baran, Bahçeşehir Üniversitesi, Istanbul, Turkey; Kara, Ömer Ekmel, Bahçeşehir Üniversitesi, Istanbul, Turkey; Hemond, Elizabeth M., Bahçeşehir Üniversitesi, Istanbul, Turkey; Tezel, Ulaş, Boğaziçi Üniversitesi, Bebek, TurkeyHeavy metal contamination in drinking water and water resources is one of the problems generated by increasing water demand and growing industrialization. Heavy metals can be toxic to humans and other living beings when their intake surpasses a certain threshold. Generally, heavy metal contamination analysis of water resources requires qualified experts with specialized equipment. In this paper, we introduce a method for citizen-based water-quality monitoring through simple pattern classification of water crystallization using a smartphone and portable microscope. This work is a first step toward the development of a Water Expert System smartphone application that will provide the ability to analyze water resource contamination remotely by sending images to the database and receiving an automatic analysis of the sample via machine learning software. In this study, we show the ability of the method to detect Fe 2 mg/1 L, 5 mg/L,10 mg/L polluted distilled water compared with other heavy metals (Al, Pb) pollution. The experimental results show that the classification used method has an accuracy greater than 90%. © 2022 Elsevier B.V., All rights reserved.Publication Metadata only Microservice Interaction Prediction in Communication Platform as a Service, Hizmet Olarak Iletişim Platformunda Mikroservis Etkileşim Tahmini(Institute of Electrical and Electronics Engineers Inc., 2022) Aktaş, Kemal; Kilinc, H. Hakan; Arica, Nafiz; Aktaş, Kemal, Bilgisayar Mühendisliǧi Bölümü, Bahçeşehir Üniversitesi, Istanbul, Turkey; Kilinc, H. Hakan, Ar-Ge Merkezi, Istanbul, Turkey; Arica, Nafiz, Bilgisayar Mühendisliǧi Bölümü, Bahçeşehir Üniversitesi, Istanbul, TurkeyIn telecommunication platforms, it is necessary to monitor the system, take quick action against possible and observed errors, and ensure the continuous operability of the system. However, debugging and problem addressing processes can take a long time in microservice architecture-based platforms with high number of users and heavy traffic. This difficulty arises from the necessity of detecting microservice interactions. In the first phase of this study, it has been observed that providing microservice flows and interactions saves operational teams a significant time during the debugging process. In this scope, different machine learning-based models that predict microservice interactions have been developed and performances are compared. In these models, log patterns are extracted on microservice log data and the interaction map of the mentioned microservice is created by estimating the previous and next microservices that the current microservice interacts with at a certain moment. In the experiments on microservice logs working with the basic call scenario, successful estimation results were obtained that could contribute positively to the debugging process. © 2022 Elsevier B.V., All rights reserved.
- «
- 1 (current)
- 2
- 3
- »