Çalım, Ali | Özer, Mahmut | Uzuntarla, Muhammet

Proceedings | 2018 | 26th IEEE Signal Processing and Communications Applications Conference, SIU 2018 , pp.1 - 4

Experimental and theoretical studies in recent years suggest that the first spike latency is an effective information carrier and contains more neural information than other spikes. Noise Delayed Decay (NDD) phenomenon emerges when the first spike latency of the neuron exposed to the periodic driving is maximum at a certain noise intensity interval. In this study, the latency dynamics of a single Hodgkin-Huxley neuron is investigated under periodic driving, background activity through dynamic synapses, and channel noise. The system response with first spike latency is investigated as a function of the presynaptic firing rate, the pa . . .rameter with an appropriate biophysical reality to control the level of activity in the nervous system. First, NDD behavior is investigated under suprathreshold stimulation in the presence of synapses at different levels of depression and channel noise. It is then desired to observe the NDD phenomenon under subthreshold stimulation with the same strategy. Our results have shown that the background activity occurring in the presence of dynamic synapses and the channel noise are significant system dynamics in observing the NDD behavior. © 2018 IEEE Daha fazlası Daha az

Kaya, Ceren | Erkaymaz, Okan | Ayar, Orhan | Özer, Mahmut

Proceedings | 2018 | 26th IEEE Signal Processing and Communications Applications Conference, SIU 2018 , pp.1 - 4

Refractive disorders are common health problems in the community and they are the most important cause of visual impairment. In this study, it was aimed to classify the individuals who have hypermetropia and myopia refractive disorders or not. For this, horizontal and vertical Electrooculogram (EOG) signal data from the right and left eyes of the individuals were used. The performance of the data was investigated by using Logistic Regression (LR), Naive Bayes (NB), Random Forest (RF) and REP Tree (RT) data mining methods. According to the obtained results, REP Tree method has shown the most successful classification performance to d . . .etect hypermetropia and myopia refractive disorders from Electrooculogram (EOG) signals. © 2018 IEEE Daha fazlası Daha az

Erkan, Yasemin | Saraç, Zehra | Yılmaz, Ergin

Proceedings | 2018 | 26th IEEE Signal Processing and Communications Applications Conference, SIU 2018 , pp.1 - 4

In this study, the effects of electrically non-excitable astrocyte cell on the weak signal detection capacity of Hodgkin-Huxley neuron are investigated. To do this, by applying a subthreshold weak signal to neuron, we investigate the weak signal detection capacity of the neuron depending on calcium channel noise stemmed from random open-close fluctuations of calcium channels. Obtained results show that astrocyte decreases the weak signal detection capacity of Hodgkin-Huxley neuron. © 2018 IEEE.

Erkaymaz, Okan | Palabaş, Tuğba

Proceedings | 2018 | 26th IEEE Signal Processing and Communications Applications Conference, SIU 2018 , pp.1 - 4

Computer assisted automatic diagnostic systems are used for the purpose of speeding up diagnosis and treatment and helping to make the right decision. In this study, cervical cancer is identified using four basic classifiers: Naive Bayes (NB), k-Nearest Neighbor (kNN), Multilayer Perceptron (MLP) and Decision Trees (KA-C4.5) algorithms and random subspaces ensemble algorithm. Gain Ratio Attribute Evaluation (GRAE) feature extraction algorithm is applied to contribute to classification performance. The classification results obtained with all datasets and reduced datasets are compared with respect to performance criteria such as accu . . .racy, Root Mean Square Error (RMSE), Sensitivity, Specificity performance criteria. According to the obtained performance analysis, it is seen that the classification performance with the random subspace ensemble algorithm using the kNN basic classifier on the reduced data set is the highest (%95.51). © 2018 IEEE Daha fazlası Daha az

Uzun, Rukiye | İşler, Yalçın | Erkaymaz, Okan | Kocadayı, Yasemin

Proceedings | 2018 | 26th IEEE Signal Processing and Communications Applications Conference, SIU 2018 , pp.1 - 4

In this study, a clustering algorithm based on adaptive neural fuzzy inference system (ANFIS) was used for computer-assisted diagnosis of the vertebral column disorder from machine learning databases of UCI (University of California Irvine). Features of pelvic incidence, pelvic tilt and lumbar lordosis angle given in this dataset was applied to the inputs of the algorithm. The performance of algorithm was evaluated using mean square error and regression coefficient criteria to discriminate patients with vertebral column disease from healthy subjects. As a result, the classification performance of 90.82% was obtained by using the gen . . .erated model of ANFIS. © 2018 IEEE Daha fazlası Daha az

Baysal, Veli | Yılmaz, Ergin

Proceedings | 2018 | 26th IEEE Signal Processing and Communications Applications Conference, SIU 2018 , pp.1 - 4

In this paper, effects of autapse on signal detection capacity of Izhikevich neuron in feed-forward-loop network motifs are investigated. Obtained results showed that autapse significantly enchances singal detection of Izhikevich neuron at proper autaptic time delay values compared without autapse. Also, it is seen that feed-forward-loop motifs have significant effects on signal detection ability of Izhikevich neuron. It is obtained that signal detection of Izhikevich neuron are best in T1 feed-forward-loop motif. © 2018 IEEE.

Çalım, Ali | Özer, Mahmut | Uzuntarla, Muhammet

Proceedings | 2018 | 26th IEEE Signal Processing and Communications Applications Conference, SIU 2018 , pp.1 - 4

In this study, effects of the heterogeneity in neuronal networks and subthreshold signal features on subthreshold signal detection in the nervous system is investigated. As most of studies in the literature investigate the subject by considering neuron populations as homogenous systems, in this study, the populations are considered as heterogeneous in terms of neuronal excitability. The information processing performance of the neuron populations is systematically studied by using mathematical equations for modeling the dynamics of the neurons, which are basic units of the system. As a result of the simulations performed, it is seen . . . that the sub-threshold signal frequency and the heterogeneity in the excitability are important system parameters for optimizing the information encoding performance. It is shown that the population encoding performance is maximized depending on the subthreshold signal frequency at different optimum levels of heterogeneity in the population. © 2018 IEEE Daha fazlası Daha az

Altan A. | Hacioglu R.

Conference Object | 2018 | 26th IEEE Signal Processing and Communications Applications Conference, SIU 2018 , pp.1 - 4

In this study, focuses on the non-linear Hammerstein model under external disturbance with white Gaussian noise based on the experimental input (motor velocities) and output (end effector position) data of the three axes gimbal system on the Unmanned Aerial Vehicle (UAV), which is autonomously moving for route tracking. The performance of UAV in reaching the target point on a planned route in sinusoidal form in avoiding obstacles, depends on the route tracking performance of the three axes gimbal system on the UAV. In intelligence activities such as exploration and surveillance, Hammerstein and Nonlinear AutoRegressive and Moving Av . . .erage (NARMA) models of the gimbal system with great importance for UAV that real time image transmission and in the tasks of leaving payloads to targets with unknown coordinates with the least mistake are obtained. The parameters of the obtained models are estimated by Recursive Least Squares (RLS) algorithm and the model performances are compared. © 2018 IEEE Daha fazlası Daha az

Karasu S. | Sarac Z.

Conference Object | 2018 | 26th IEEE Signal Processing and Communications Applications Conference, SIU 2018 , pp.1 - 4

In this study, instantaneous envelope, phase and frequency series are obtained by Hilbert transform for Power Quality (PQ-Power Quality) disturbances signals. Rms, Thd, energy, entropy and statistical properties are applied to these series. With the wrapper feature selection approach, a set of features is obtained that has a small number of feature subset and a high performance from 36 features. Genetic Algorithm (GA) is used as a search algorithm and the classifier algorithm is K nearest neighborhood (KNN). Support Vector Machines (SVM) for selected features are also used in the classification step. The learning algorithm is obtain . . .ed as KNN, the model performance that classifies PQ classes with 99.07%. The number of feature sets is 8. In addition, performance under noisy data is also tested to show that the generated model has a generalized structure. © 2018 IEEE Daha fazlası Daha az

Kaya G.U. | Sarac Z.

Conference Object | 2018 | 26th IEEE Signal Processing and Communications Applications Conference, SIU 2018 , pp.1 - 4

The purpose of this study is to train the reconstructed sound waves, which is obtained from recording holograms via digital holography, and reduce the noise from this sound wave without using noise Altering techniques. The noise reduction is achieved by approximating the reconstructed sound wave trained by YSA to the actaul recording sound wave. A network topology is created for training and the system was tested. The performance of the system is given by showing how closely the sound wave trained by YSA approaches the actual sound wave. © 2018 IEEE.

Karasu S. | Altan A. | Sarac Z. | Hacioglu R.

Conference Object | 2018 | 26th IEEE Signal Processing and Communications Applications Conference, SIU 2018 , pp.1 - 4

In this study, Bitcoin prediction is performed with Linear Regression (LR) and Support Vector Machine (SVM) from machine learning methods by using time series consisting of daily Bitcoin closing prices between 2012-2018. The prediction model with include the least error is obtained by testing with different parameter combinations such as SVM with including linear and polynomial kernel functions. Filters with different weight coefficients are used for different window lengths. For different window lengths, Bitcoin price prediction is made using filters with different weight coefficients. 10-fold cross-validation method in training ph . . .ase is used in order to construct a model with high performance independent of the data set. The performance of the obtained model is measured by means of statistical indicators such as Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Pearson Correlation. It is seen that the price prediction performance of the proposed SVM model for Bitcoin data set is higher than that of the LR model. © 2018 IEEE Daha fazlası Daha az

Kara F. | Kaya H.

Conference Object | 2018 | 26th IEEE Signal Processing and Communications Applications Conference, SIU 2018 , pp.1 - 4

NOMA is as a strong candidate for the Future Radio Access Network (FRA) due to its potential to support massive connectivity and high spectral efficiency. However, the most important drawback of NOMA is the error during Successive Interference Canceller (SIC) is implemented because of the inter-user interferences. In this paper, we derive closed-form exact Bit-Error Rate expressions for Downlink(DL) NOMA over Nakagami-m fading channels in the presence of SIC errors. The derived expressions are validated by the computer simulations. It is shown that the m parameter still represents the diversity order like as OMA systems. Besides, th . . .e BER performances of users for NOMA have substantially depended on the power allocation coefficient. © 2018 IEEE Daha fazlası Daha az