Yazar "Şahin, Murat" seçeneğine göre listele

Listeleniyor 1 - 4 / 4
  • Küçük Resim
    Öğe
    Classification of electrocardiogram (ECG) data using deep learning methods
    (Institute of Electrical and Electronics Engineers Inc., 2020) Bozyiğit, Fatma; Erdemir, Fatih; Şahin, Murat; Kılınç, Deniz
    Classification is one of the most widely used techniques in healthcare, especially concerning diagnosing cardiac disorders. Arrhythmia is a disorder of the heartbeat rate or rhythm, which may occur sporadically in daily life. Electrocardiogram (ECG) is an important diagnostic tool for analysing cardiac tissues and structures. It includes information about the heart structure and the function of its electrical conduction system. Since manual analysis of heartbeat rate is time-consuming and prone to errors, automatic recognition of arrhythmias using ECG signals has become an increasingly popular research focus in recent years. Current ECG analysis systems in literature generally have implemented well known machine learning algorithms. Due to the advent of powerful parallel computing hardware and the big data technologies, deep learning has also become a widely preferred technique in healthcare applications. In our study, we use ECG data in MIT-BIH Arrhythmia Database to develop a Convolutional Neural Networks (CNN) which is a deep feed-forward neural network type. The parameter tuned/optimized version of the proposed algorithm on top of the reduced feature dimension is more efficient than state of the art in terms of accuracy. Finally, we also compare the results of the proposed algorithm with Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), and we provide the corresponding results in related sections. © 2020 IEEE.
  • Küçük Resim
    Öğe
    Comparison of tabular data generation algorithms using gans
    (İzmir Bakırçay Üniversitesi Lisansüstü Eğitim Enstitüsü, 2023) Şahin, Murat; Alpkoçak, Adil
    Veri mahremiyeti, eksik veri, tutarsız veri ve veri yetersizliği gibi çeşitli veri kaynaklı sorunlar nedeniyle sentetik verilere olan ihtiyaç her geçen gün artmaktadır. Sentetik veri üretmek için GAN (Generative Adversarial Network), VAE (Variational Autoencoders), kopula vb. birçok yaklaşım geliştirilmiştir. Bu tezde farklı mimarilere ve özelliklere sahip 5 GAN modeli (C-GAN, CT-GAN, DAT-GAN, DPC-GAN, Copula-GAN) 5 iyi bilinen veri seti (adult, credit, census, insurance, cardio-train) üzerinde deneylerle karşılaştırdık. Değerlendirme sırasında sentetik verilerin gerçek verilere ne kadar benzediğini anlamak için çeşitli istatistiksel testler uyguladık ve elde edilen sonuçları paylaştık. Ayrıca, makine öğrenimi modellerini kullanarak gerçek veriler yerine sentetik verilerin nasıl kullanılabileceğini de gösterdik. Karışık veri türleri, uzun kuyruk problemleri, normal olmayan (çarpık) dağılımlar, çok tepeli dağılımlar, seyrek kodlanmış vektörler, yüksek derecede dengesiz kategorik sütunlar gibi bazı açık sorunların deneylerimizde nasıl sonuç verdiğini grafiklerle sunduk. Deneysel sonuçlara bakıldığında kullanılacak GAN modelinin başarısı veri setlerinin özelliklerine ve boyutlarına göre değişmekle birlikte deneydeki en tutarlı ve başarılı sonuçlar Copula-GAN, DPC-GAN ve CT-GAN modellerine ait gözlemlenmiştir.
  • Küçük Resim
    Öğe
    A Hybrid Stock optimization Approach for Inventory Management
    (Institute of Electrical and Electronics Engineers Inc., 2021) Çimen, Egemen Berki; Kurban, İlknur; Selmanoğlu, Özgür; Şahin, Murat; Kılınç, Deniz
    Modern world is rapidly evolving around knowledge and business that know how to use knowledge shows superiority. Thus, smart decision making becomes vital for the modern business world to achieve sustainability in life and business. Especially, with a world of scarce resources, utilizing knowledge would play critical not only today but also for future. Moreover, using knowledge is inevitable in supply chain and inventory management must be supported with smart algorithms and modern heuristics to avoid excessive inventory while fighting with stockout. Therefore, this study explores the opportunity for inventory planning with heuristics and tailored techniques as well as how to hybridize modern heuristics and tailored techniques. In this study, inventory optimization experiments are proposed to model spare part inventory management and find the best way to determine reorder amount to deal with shortage and excessive inventory. Four heuristics which are i) basic golden ratio, ii) simulated annealing, iii) statistical rule-based heuristic and iv) hybrid algorithm of simulated annealing and statistical rule-based heuristic are evaluated on existing spare part dataset with an example from real-life part supplier named Eldor. In this test case, 400 products are analyzed, and best reorder points and amounts are selected with the help of heuristics. Heuristics' main structures and parameters are adjusted for the problem's need and improvement on quality of results. Parameters are determined according to trial and error with experts' guidance on heuristics. The best result suggests 8% improvement on cost and 37% improvement on inventory load could achieve with the help of heuristics. These solutions are usable against even hard constraints like shortage. © 2021 IEEE.
  • [ X ]
    Öğe
    REGRESSION BASED RISK ANALYSIS IN LIFE INSURANCE INDUSTRY
    (Ahmet Ali SÜZEN, 2020) Bozyiğit, Fatma; Şahin, Murat; Gündüz, Tolga; Işık, Cem; Kılınç, Deniz
    Risk analysis is a crucial part for classifying applicants in life insurance business. Since the traditional underwriting strategies are time-consuming, recent works have focused on machine learning based methods to make the steps of underwriting more effective and strengthening the supervisory. The aim of this study is to evaluate the linear and non-linear regression-based models to determine the degree of risk. Therefore, four linear and non-linear regression algorithms are trained and evaluated on a life insurance dataset. The parameters of algorithms are optimized using Grid Search approach. The experimental results show that the non-linear regression models achieve more accurate predictions than linear regression models and the LGBM algorithm has the best performance among the all regression models with the highest R2, lowest MAE and RMSE values.