Yazar "Ozbay, Yueksel" seçeneğine göre listele
Listeleniyor 1 - 4 / 4
Sayfa Başına Sonuç
Sıralama seçenekleri
Öğe Classification of ECG Arrhythmias using Type-2 Fuzzy Clustering Neural Network(IEEE, 2009) Ceylan, Rahime; Ozbay, Yueksel; Karlik, BekirIn this study, Type-2 Fuzzy Clustering Neural Network (T2FCNN) architecture realized for classification of electrocardiography arrhythmias is presented. Type-2 fuzzy clustering neural network is cascade structure formed by clustering and classification stages. In T2FCNN architecture, clustering stage consisted of select best patterns in all patterns that belongs to same class is executed by type-2 fuzzy c-means clustering (T2FCM). The aim of using T2FCM clustering algorithm is to reduce classification error of neural network by optimization of training pattern set. A new training set consisted of cluster centers obtained by type-2 fuzzy c-means clustering algorithm for each class as separately is formed inputs of neural network. Neural network is trained using backpropagation algorithm. Proposed structure is used classification of five ECG signal class composed normal sinus rhythm, sinus bradycardia, sinus arrhythmia, right bundle branch block and left bundle branch block. Data used in this study is obtained from Physionet database, that belongs to MIT-BIH ECG Arrhythmia Database. In the end of making applications, proposed T2FCNN structure is classified ECG arrhythmias with 99% detection rate.Öğe Effects of Complex Wavelet Transform with Different Levels in Classification of ECG Arrhytmias using Complex-Valued ANN(IEEE, 2009) Ceylan, Murat; Ozbay, YuekselIn this study, a new structure formed by complex wavelet transform (CHIT) with different levels and complex-valued artificial neural network (CVANN) is proposed for classification of ECG arryhytmias. In this structure, features of ECG data are extracted using CWT and data size is reduced. After then, four statistical features (maximum value, minimum value, mean value and standard deviation) are obtained from; extracted features. These new statistical features are presented to CVANN as inputs. Data set used in this studs,, including five different arrhytmias (normal sinus rhythm, right bundle branch block, left bundle branch block, atrial fibrilation and atrial flutter), are selected from MIT-BIH ECG database. Number of samples in training and test sets for each pattern is reduced from 200 real-valued samples to 100, 50 and 25 complex-valued samples using first level CWT second level CWT and third level CWT respectively. Classificaton results shown that arrhytmias are classified with 100 % accuracy rate using CWT with third level. Classification process was done in 32.62 second.Öğe A new approach for classification of EEG signals(IEEE, 2007) Tezel, Guelay; Ozbay, YuekselThis study presents a comparative study of the classification accuracy and speed of performance of epileptic Electroensefalogram (EEG) signals using a traditional neural network architecture based on backpropagation training algorithm, and a new neural network. The proposed network is called adaptive neural network with activation function (AAF-NN) in which adjustable parameters, It is used two different activation functions for developed study. One of theese adaptive activation functions is sigmoid function with free parameters and the other one is sum of sinusoidal function with free parameters and sigmoid function with free parameters. The adaptive activation function with free parameters is used in the hidden layer for the proposed structures based on the feed-forward neural network Experimental results have revealed that neural network with adaptive activation function is more suitable for classification EEG signals and training speed is much faster than traditional neural network with fixed sigmoid activation function.Öğe Prediction of force reduction factor (R) of prefabricated industrial buildings using neural networks(TECHNO-PRESS, 2007) Arslan, M. Hakan; Ceylan, Murat; Kaltakci, M. Yasar; Ozbay, YuekselThe force (load) reduction factor, R, which is one of the most important parameters in earthquake load calculation, is independent of the dimensions of the structure but is defined on the basis of the load bearing system of the structure as defined in earthquake codes. Significant damages and failures were experienced on prefabricated reinforced concrete structures during the last three major earthquakes in Turkey (Adana 1998, Kocaeli 1999, Duzce 1999) and the experts are still discussing the main reasons of those failures. Most of them agreed that they resulted mainly from the earthquake force reduction factor, R that is incorrectly selected during design processes, in addition to all other detailing errors. Thus this wide spread damages caused by the earthquake to prefabricated structures aroused suspicion about the correctness of the R coefficient recommended in the current Turkish Earthquake Codes (TEC - 98). In this study, an attempt was made for an approximate determination of R coefficient for widely utilized prefabricated structure types (single-floor single-span) with variable dimensions. According to the selecting variable dimensions, 140 sample frames were computed using pushover analysis. The force reduction factor R was calculated by load-displacement curves obtained pushover analysis for each frame. Then, formulated artificial neural network method was trained by using 107 of the 140 sample frames. For the training various algorithms were used. The method was applied and used for the prediction of the R rest 33 frames with about 92% accuracy. The paper also aims at proposing the authorities to change the R coefficient values predicted in TEC - 98 for prefabricated concrete structures.
