Yazar "Nour, Majid" seçeneğine göre listele
Listeleniyor 1 - 20 / 26
Sayfa Başına Sonuç
Sıralama seçenekleri
Öğe An adaptive deep learning framework to classify unknown composite power quality event using known single power quality events(Pergamon-Elsevier Science Ltd, 2021) Sindi, Hatem; Nour, Majid; Rawa, Muhyaddin; Öztürk, Şaban; Polat, KemalDistributed generation (DG) sources are preferred to meet today's energy needs effectively. The addition of many different types of renewable energy sources to the grid causes various problems in signal quality. Detection and classification of these problems increase efficiency by both the producer and the consumer. In the literature, incredibly singular and some composite power quality disturbance (PQD) detection is performed effectively. However, the multitude of composite PQD variations degrades the performance of existing algorithms. In this study, the classification of all PQD variations that may occur is performed by using singular PQD and some composite PQD signals. A different number of subcomponents representing the signal are created according to each signal characteristic. Instantaneous energies from these subcomponents are used as deep learning (DL) input. Deep learning cycles are created as much as the instantaneous energy number of each signal. Each cycle has specific features of defining a single event. Therefore, the proposed approach is able to classify composite PQD signals that it has not encountered before. The proposed method's performance is first evaluated with the known PQD events and compared with the current state-of-the-art methods in the literature. Then, a dataset containing the combinations of different events not encountered during the training is created, and the performance is evaluated on this dataset. In the experiments performed, it is revealed that the proposed framework produces higher performance than other state-of-the-art methodsÖğe Automatic classification of hypertension types based on personal features by machine learning algorithms(Hindawi Ltd, 2020) Nour, Majid; Polat, KemalHypertension (high blood pressure) is an important disease seen among the public, and early detection of hypertension is significant for early treatment. Hypertension is depicted as systolic blood pressure higher than 140 mmHg or diastolic blood pressure higher than 90 mmHg. In this paper, in order to detect the hypertension types based on the personal information and features, four machine learning (ML) methods including C4.5 decision tree classifier (DTC), random forest, linear discriminant analysis (LDA), and linear support vector machine (LSVM) have been used and then compared with each other. In the literature, we have first carried out the classification of hypertension types using classification algorithms based on personal data. To further explain the variability of the classifier type, four different classifier algorithms were selected for solving this problem. In the hypertension dataset, there are eight features including sex, age, height (cm), weight (kg), systolic blood pressure (mmHg), diastolic blood pressure (mmHg), heart rate (bpm), and BMI (kg/m(2)) to explain the hypertension status and then there are four classes comprising the normal (healthy), prehypertension, stage-1 hypertension, and stage-2 hypertension. In the classification of the hypertension dataset, the obtained classification accuracies are 99.5%, 99.5%, 96.3%, and 92.7% using the C4.5 decision tree classifier, random forest, LDA, and LSVM. The obtained results have shown that ML methods could be confidently used in the automatic determination of the hypertension types.Öğe Automatic classification of hypertension types based on personal features by machine learning algorithms (vol 2020, 2742781, 2020)(Hindawi Ltd, 2020) Nour, Majid; Polat, KemalCorrection/Düzeltme: In the article titled “Automatic Classification of Hypertension Types Based on Personal Features by Machine Learning Algorithms” [1], there was an error in the reference provided for the PPG-BP database, which can be correctly accessed atÖğe Diagnosis and classification of Parkinson's disease using ensemble learning and 1D-PDCovNN(Pergamon-Elsevier Science Ltd, 2023) Nour, Majid; Şentürk, Ümit; Polat, KemalIn this paper, we proposed a novel approach to diagnose and classify Parkinson's Disease (PD) using ensemble learning and 1D-PDCovNN, a novel deep learning technique. PD is a neurodegenerative disorder; early detection and correct classification are essential for better disease management. The primary aim of this study is to develop a robust approach to diagnosing and classifying PD using EEG signals. As the dataset, we have used the San Diego Resting State EEG dataset to evaluate our proposed method. The proposed method mainly consists of three stages. In the first stage, the Independent Component Analysis (ICA) method has been used as the pre-processing method to filter out the blink noises from the EEG signals. Also, the effect of the band showing motor cortex activity in the 7-30 Hz frequency band of EEG signals in diagnosing and classifying Parkinson's disease from EEG signals has been investigated. In the second stage, the Common Spatial Pattern (CSP) method has been used as the feature extraction to extract useful information from EEG signals. Finally, an ensemble learning approach, Dynamic Classifier Selection (DCS) in Modified Local Accuracy (MLA), has been employed in the third stage, consisting of seven different classifiers. As the classifier method, DCS in MLA, XGBoost, and 1D-PDCovNN classifier has been used to classify the EEG signals as the PD and healthy control (HC). We first used dynamic classifier selection to diagnose and classify Parkinson's disease (PD) from EEG signals, and promising results have been obtained. The performance of the proposed approach has been evaluated using the classification accuracy, F-1 score, kappa score, Jaccard score, ROC curve, recall, and precision values in the classification of PD with the proposed models. In the classification of PD, the combination of DCS in MLA achieved an accuracy of 99,31%. The results of this study demonstrate that the proposed approach can be used as a reliable tool for early diagnosis and classification of PD.Öğe A different sleep apnea classification system with neural network based on the acceleration signals(Elsevier Sci Ltd, 2020) Yüzer, Ahmet Hayrettin; Sümbül, Harun; Nour, Majid; Polat, KemalBackground and objective: The apnea syndrome is characterized by an abnormal breath pause or reduction in the airflow during sleep. It is reported in the literature that it affects 2% of middle-aged women and 4% of middle-aged men, approximately. This study has vital importance, especially for the elderly, the disabled, and pediatric sleep apnea patients. Methods: In this study, a new diagnostic method is developed to detect the apnea event by using a microelectromechanical system (MEMS) based acceleration sensor. It records the value of acceleration by measuring the movements of the diaphragm in three axes during the respiratory. The measurements are carried out simultaneously, a medical spirometer (Fukuda Sangyo), to test the validity of measurement results. An artificial neural network model was designed to determine the apnea event. For the number of neurons in the hidden layer, 1-3-5-10-18-20-25 values were tried, and the network with three hidden neurons giving the most suitable result was selected. In the designed ANN, three layers were formed that three neurons in the hidden layer, the two neurons at the input, and two neurons at the output layer. Results: A study group was formed of 5 patients (having different characteristics (age, height, and body weight)). The patients in the study group have sleep apnea (SA) in different grades. Several 12.723 acceleration data (ACC) in the XYZ-axis from 5 different patients are recorded for apnea event training and detection. The measured accelerometer (ACC) data from one of the patients (called H1) are used to train an ANN. During the training phase, MSE is used to calculate the fitness value of the apnea event. Then Apnea event is detected successfully for the other patients by using ANN trained only with H1's ACC data. Conclusions: The sleep apnea event detection system is presented by using ANN from directly acceleration values. Measurements are performed by the MEMS-based accelerometer and Industrial Spirometer simultaneously. A total of 12723 acceleration data is measured from 5 different patients. The best result in 7000 iterations was reached (the number of iterations was tried up to 10.000 with 1000 steps). 605 data of only H1 measurements are used to train ANN, and then all data used to check the performance of the ANN as well as H2, H3, H4, and H5 measurement results. MSE performance benchmark shows us that trained ANN successfully detects apnea events. One of the contributions of this study to literature is that only ACC data are used in the ANN training step. After training for one patient, the ANN system can monitor the apnea event situation on-line for others. (C) 2020 Elsevier Ltd. All rights reserved.Öğe Epileptic seizure detection based on new hybrid models with electroencephalogram signals(Elsevier Science Inc, 2020) Polat, Kemal; Nour, MajidObjectives: Epileptic seizures are one of the most common diseases in society and difficult to detect. In this study, a new method was proposed to automatically detect and classify epileptic seizures from EEG (Electroencephalography) signals. Methods: In the proposed method, EEG signals classification five-classes including the cases of eyes open, eyes closed, healthy, from the tumor region, an epileptic seizure, has been carried out by using the support vector machine (SVM) and the normalization methods comprising the z-score, minimum-maximum, and MAD normalizations. To classify the EEG signals, the support vector machine classifiers having different kernel functions, including Linear, Cubic, and Medium Gaussian, have been used. In order to evaluate the performance of the proposed hybrid models, the confusion matrix, ROC curves, and classification accuracy have been used. The used SVM models are Linear SVM, Cubic SVM, and Medium Gaussian SVM. Results: Without the normalizations, the obtained classification accuracies are 76.90%, 82.40%, and 81.70% using Linear SVM, Cubic SVM, and Medium Gaussian SVM, respectively. After applying the z-score normalization to the multi-class EEG signals dataset, the obtained classification accuracies are 77.10%, 82.30%, and 81.70% using Linear SVM, Cubic SVM, and Medium Gaussian SVM, respectively. With the minimum-maximum normalization, the obtained classification accuracies are 77.20%, 82.40%, and 81.50% using Linear SVM, Cubic SVM, and Medium Gaussian SVM, respectively. Moreover, finally, after applying the MAD normalization to the multi-class EEG signals dataset, the obtained classification accuracies are 76.70%, 82.50%, and 81.40% using Linear SVM, Cubic SVM, and Medium Gaussian SVM, respectively. Conclusion: The obtained results have shown that the best hybrid model is the combination of cubic SVM and MAD normalization in the classification of EEG signals classification five-classes. (C) 2020 AGBM. Published by Elsevier Masson SAS. All rights reserved.Öğe A healthcare evaluation system based on automated weighted indicators with cross-indicators based learning approach in terms of energy management and cybersecurity(Elsevier Ireland Ltd, 2020) Nour, Majid; Sindi, Hatem; Abozinadah, Ehab; Öztürk, Şaban; Polat, KemalObjective: Hospital performance evaluation is vital in terms of managing hospitals and informing patients about hospital possibilities. Also, it plays a key role in planning essential issues such as electrical energy management and cybersecurity in hospitals. In addition to being able to make this measurement objectively with the help of various indicators, it can become very complicated with the participation of subjective expert thoughts in the process. Method: As a result of budget cuts in health expenditures worldwide, the necessity of using hospital resources most efficiently emerges. The most effective way to do this is to determine the evaluation criteria effectively. Machine learning (ML) is the current method to determine these criteria, determined by consulting with experts in the past. ML methods, which can remain utterly objective concerning all indicators, offer fair and reliable results quickly and automatically. Based on this idea, this study provides an automated healthcare system evaluation framework by automatically assigning weights to specific indicators. First, the ability of hands to be used as input and output is measured. Results: As a result of this measurement, indicators are divided into only input group (group A) and both input and output group (group B). In the second step, the total effect of each input on the output is calculated by using the indicators in group B as output sequentially using the random forest of the regression tree model. Conclusion: Finally, the total effect of each indicator on the healthcare system is determined. Thus, the whole system is evaluated objectively instead of a subjective evaluation based on a single output.Öğe Machine learning and electrocardiography signal-based minimum calculation time detection for blood pressure detection(Hindawi Ltd, 2022) Nour, Majid; Kandaz, Derya; Uçar, Muhammed Kürşad; Polat, Kemal; Alhudhaif, AdiObjective. Measurement and monitoring of blood pressure are of great importance for preventing diseases such as cardiovascular and stroke caused by hypertension. Therefore, there is a need for advanced artificial intelligence-based systolic and diastolic blood pressure systems with a new technological infrastructure with a noninvasive process. The study is aimed at determining the minimum ECG time required for calculating systolic and diastolic blood pressure based on the Electrocardiography (ECG) signal. Methodology. The study includes ECG recordings of five individuals taken from the IEEE database, measured during daily activity. For the study, each signal was divided into epochs of 2-4-6-8-10-12-14-16-18-20 seconds. Twenty-five features were extracted from each epoched signal. The dimension of the dataset was reduced by using Spearman's feature selection algorithm. Analysis based on metrics was carried out by applying machine learning algorithms to the obtained dataset. Gaussian process regression exponential (GPR) machine learning algorithm was preferred because it is easy to integrate into embedded systems. Results. The MAPE estimation performance values for diastolic and systolic blood pressure values for 16-second epochs were 2.44 mmHg and 1.92 mmHg, respectively. Conclusion. According to the study results, it is evaluated that systolic and diastolic blood pressure values can be calculated with a high-performance ratio with 16-second ECG signals.Öğe The methods toward improving communication performance in transparent radio frequency signals(Hindawi Ltd, 2020) Daldal, Nihat; Nour, Majid; Polat, KemalIn wireless digital communications, amplitude-shift keying (ASK) and frequency-shift keying (FSK) modules are often used and radio frequency (RF), communication synchronization, and noise problems affect the performance very much. In particular, the sending of byte-type data called synchronous and preamble before sending data in intermodule communication increases the sent data and decreases the speed. Also, the microcontroller at the output of the RF receiver module continuously listens to the RF noise and analyzes incoming data, but this increases the processing load of the microcontroller. Moreover, it reduces the speed of performing other operations. In this study, a transparent RF transmitter and receiver have been investigated, and methods for increasing the communication performance of the modules have been proposed and performed. Two of the proposed methods prevented the continuous listening of the microprocessor in the RF receiver structure so that the microprocessor can be used with other processes. In other methods, the compression of the data size was achieved because the transmission of a series of data in RF communication systems was limited to a certain extent. In the last section of the study, since the RF modules have failed to transmit the data due to corruption in the extended data dimensions, the bit carrier control security code has been created for the data series and more healthy communication has been performed.Öğe Morphology properties of scapular spine relative to reverse shoulder arthroplasty: A biomechanical study(Elsevier Science Ltd, 2023) Chen, Junfeng; Chen, Chen; Nour, Majid; Liu, Debao; Zhu, Youyu; Polat, KemalObjective: Stress fractures of the scapular spine are often a consequence of reverse shoulder arthroplasty (RSA) reconstruction. Knowledge of the morphology and strength of the scapular spine may have potential value preoperative planning of RSA. However, the morphometric and biomechanics of the scapular spine have not been thoroughly investigated yet. This study aimed to evaluate the association between different morphology of the scapular spine and biomechanics to identify scapular spine with weaker strength that may be destroyed after RSA.Methods: Anatomical morphology of 102 dry scapulae was studied. Nine bony landmarks associated with the anatomical morphology of the scapular spine were measured. After the measurements, each specimen was loaded to failure under an axial compressive load. All the samples were analyzed at the point of failure to calculate the failure load and energy.Results: Five types of scapular spine were observed in the population. Type I (fusiform shape) was the most frequent (32.4%), and the least common type included type V (S-shape) (4.9%). The morphological parameters for type II (slender rod shape) were smaller than those for the other types. Type II scapular spine withstood the lowest failure load (381.87 +/- 25.31 N) and energy (1.45 +/- 0.56 J), followed by type V (530.98 +/- 25.26 N and 1.75 +/- 0.49 J).Conclusion: Significant variation was observed in the anatomical morphology of scapular spine. Slender rod shaped and S-shaped scapular spine bore lower load and energy, which may be the potential anatomical related factors of stress fractures after RSA.Öğe New approaches to epileptic seizure prediction based on EEG signals using hybrid CNNs(Inderscience Enterprises Ltd, 2024) Nour, Majid; Arabaci, Bahadir; Ocal, Hakan; Polat, KemalThis study employs the University of Bonn Dataset to address the importance of frequency information in EEG data and introduces a methodology utilising the short-time Fourier transform. The proposed method transforms conventional 1D EEG signals into informative 2D spectrograms, offering an approach for advancing the detection of neurological diseases. Integrating advanced CNN architectures with the conversion of EEG signals into 2D spectrograms forms the foundation of our proposed methodology. The 1D CNN model utilised in this study demonstrates exceptional performance metrics, achieving a specificity of 0.996, an overall test accuracy of 0.991, a sensitivity of 0.987, and an F1 score of 0.989. Shifting to the 2D approach discloses a slight reduction in accuracy to 0.987, sensitivity of 0.976, specificity of 0.988, and an F1 score of 0.97. This analysis provides nuanced insights into the performance of 1D and 2D CNNs, clarifying respective strengths in the context of neurological disease detection.Öğe A new generation communication system based on deep learning methods for the process of modulation and demodulation from the modulated images(Hindawi Ltd, 2022) Daldal, Nihat; Sezer, Zeynel Abidin; Nour, Majid; Alhudhaif, Adi; Polat, KemalDemodulating the modulated signals used in digital communication on the receiver side is necessary in terms of communication. The currently used systems are systems with a variety of hardware. These systems are used separately for each type of communication signal. A single algorithm facilitates the classification and subsequent demodulation of signals without needing hardware instead of extra hardware cost and complex systems. This study, which aims to make modulation classification by using images of signals, provides this convenience. In this study, a classification and demodulation process is done by using images of digital modulation signals. Convolutional neural network (CNN), a deep learning algorithm, has been used for classification and recognition. Images of the signals of quadrate amplitude shift keying (QASK), quadrate frequency shift keying (QFSK), and quadrate phase shift keying (QPSK) digital modulation types at noise levels of 0 dB, 5 dB, 10 dB, and 15 dB were used. Thanks to this algorithm, which works without hardware, the success achieved is around 98%. Python programming language and libraries have been used in training and testing the algorithm. Demodulation processes of these signals have been performed for demodulation using the nonlinear autoregressive network with exogenous inputs (NARX) algorithm, an artificial neural network. As a result of using MATLAB, the NARX algorithm achieved approximately 94% success in obtaining the information signal. Thanks to the work done, it will be possible to classify and demodulate other communication signals without extra hardware.Öğe A novel approach to skin lesion segmentation: Multipath fusion model with fusion loss(Hindawi Ltd, 2022) Alhudhaif, Adi; Öcal, Hakan; Barışçı, Necaattin; Atacak, İsmail; Nour, Majid; Polat, KemalSegmentation of skin lesions plays a very important role in the early detection of skin cancer. However, indistinguishability due to various artifacts such as hair and contrast between normal skin and lesioned skin is an important challenge for specialist dermatologists. Computer-aided diagnostic systems using deep convolutional neural networks are gaining importance in order to cope with difficulties. This study focuses on deep learning-based fusion networks and fusion loss functions. For the automatic segmentation of skin lesions, U-Net (U-Net + ResNet 2D) with 2D residual blocks and 2D volumetric convolutional neural networks were fused for the first time in this study. Also, a new fusion loss function is proposed by combining Dice Loss (DL) and Focal Tversky Loss (FTL) to make the proposed fused model more robust. Of the 2594 image dataset, 20% is reserved for test data and 80% for training data. In test data training, a Jaccard score of 0.837 and a dice score of 0.918 were obtained. The proposed model was also scored on the ISIC 2018 Task 1 test images, whose ground truths were not shared. The proposed model performed well and achieved a Jaccard index of 0.800 and a dice score of 0.880 in the ISIC 2018 Task 1 test set. In addition, it has been observed that the new fused loss function obtained by fusing Focal Tversky Loss and Dice Loss functions in the proposed model increases the robustness of the model in the tests. The proposed new loss function fusion model has outstripped the cutting-edge approaches in the literature.Öğe A novel classification framework using multiple bandwidth method with optimized CNN for brain-computer interfaces with EEG-fNIRS signals(SPRINGER LONDON LTD, 2021) Nour, Majid; Öztürk, Şaban; Polat, KemalThe most effective way to communicate between the brain and electronic devices in the outside world is the brain-computer interface (BCI) systems. BCI systems use signals of being through neural activity in the brain to fulfill this function. Traditional BCI systems use electroencephalography (E.E.G.) signals due to their characteristics, such as temporal resolution, cost, and noninvasive nature. However, the inherent complex features make the analysis process very difficult. In addition, its sensitivity to internal and external noise affects performance negatively. Near-infrared spectroscopy (NIRS), which describes brain hemodynamics, is a noninvasive method and robust against the problems that E.E.G. suffers. We present an effective study examining the effects of E.E.G. and NIRS signals for BCI and investigating the contribution of their combination to performance. Also, a novel classification framework using multiple bandwidth method with optimized convolution neural network (CNN) is proposed. The proposed method classifies the recorded E.E.G. and NIRS signals according to the imagination of opening and closing the subjects' right and left hands. A CNN architecture including fully connected layer optimization using E.E.G. and NIRS signals is trained in an end-to-end manner. Instead of using a single bandwidth as in the literature, multiple bandwidths are used in the training process. In this way, information loss in band filtering tasks is prevented. Performance indicators obtained from experiments performed using the proposed framework are superior to current state-of-the-art methods in the literature in the most significant performance metrics: accuracy and stability. The proposed approach has a higher classification performance than current state-of-the-art methods, with an accuracy performance of 99.85%. On the other hand, in order to test the performance of the proposed CNN method, a detailed ablation study section on single-band experiments and including analysis of each component is presented.Öğe A novel cuffless blood pressure prediction: Uncovering new features and new hybrid ml models(MDPI, 2023) Nour, Majid; Polat, Kemal; Şentürk, Ümit; Arıcan, MuratThis paper investigates new feature extraction and regression methods for predicting cuffless blood pressure from PPG signals. Cuffless blood pressure is a technology that measures blood pressure without needing a cuff. This technology can be used in various medical applications, including home health monitoring, clinical uses, and portable devices. The new feature extraction method involves extracting meaningful features (time and chaotic features) from the PPG signals in the prediction of systolic blood pressure (SBP) and diastolic blood pressure (DBP) values. These extracted features are then used as inputs to regression models, which are used to predict cuffless blood pressure. The regression model performances were evaluated using root mean squared error (RMSE), R-2, mean square error (MSE), and the mean absolute error (MAE). The obtained RMSE was 4.277 for systolic blood pressure (SBP) values using the Matern 5/2 Gaussian process regression model. The obtained RMSE was 2.303 for diastolic blood pressure (DBP) values using the rational quadratic Gaussian process regression model. The results of this study have shown that the proposed feature extraction and regression models can predict cuffless blood pressure with reasonable accuracy. This study provides a novel approach for predicting cuffless blood pressure and can be used to develop more accurate models in the future.Öğe A novel demodulation structure for quadrate modulation signals using the segmentary neural network modelling(Elsevier Sci Ltd, 2020) Daldal, Nihat; Nour, Majid; Polat, KemalIn digital communication, the baseband information signal is modulated by the high-frequency carrier to produce a passband signal and applied to the transmission line. QASK (quadrature amplitude-shift keying), QFSK (quadrature frequency-shift keying), and QPSK (quadrature phase-shift keying) modulations from Quadrate type digital modulations are used for high-speed communication in passband digital modulations. They are a widespread modulation for fast and easy data transmission, especially in wireless communication and modem devices. In these modulations, four separate carriers are used, and since each carrier represents 2 bits, the transmission rate doubles compared to conventional digital modulation. This study aims to obtain the baseband signal from quadrate type modulation signals. For this purpose, all 8-bit data between Decimal 0-255 were obtained in QASK, QFSK, and QPSK modulations and signal matrices were formed. The various SNR (signal to noise ratio) values of 5 dB-10 dB-15 dB-20 dB noise were added to these signals to examine the system performance during the modulated signal transmission. The generated signals were given to the demodulation system developed using different methods and tested. The best results were obtained in developed a segmentary NN (Neural Network). In this study, it was observed that modulation signal matrices were given directly to an ANN (Artificial Neural Network) and that the results could not be predicted with the application of noise matrices for testing. Then the modulation signals with the proposed method are divided into four parts. Each segment represents a two-bit piece of data. In the case of a column matrix, four parts were applied as input to the neural network model. In the output of ANN, the result matrix to be predicted is created. Each modulation signal applied to the network input was classified between 0 and 3 at the output. Modulation data-carrying 8 bits are applied to the network in 4 steps and classified. 4 separate classification data from the ANN output is converted back to 2-bit logic. Therefore, signals carrying 8 bits of data are obtained in 4 steps. After the formation of the ANN network, baseband digital signal estimation was performed quickly in 4 steps across each byte modulation signals under different noises coming into the network, and demodulation data was successfully achieved. (C) 2020 Elsevier Ltd. All rights reserved.Öğe A novel hybrid deep learning approach including combination of 1D power signals and 2D signal images for power quality disturbance classification(PERGAMON-ELSEVIER SCIENCE LTD, 2021) Sindi, Hatem; Nour, Majid; Rawa, Muhyaddin; Öztürk, Şaban; Polat, KemalAs a result of the widespread use of power electronic equipment and the increase in consumption, the importance of effective energy policies and the smart grid begins to increase. Nonlinear loads and other loads in electric power systems are considered as the main reason for power quality disturbance. Distortions in signal quality and shape due to power quality disturbance cause a decrease in total efficiency. The proposed hybrid convolutional neural network method consists of a 1D convolutional neural network structure and a 2D convolutional neural network structure. The features acquired by these two convolutional neural network architectures are classified using the fully connected layer, which is traditionally used as the classifier of convolutional neural network architectures. Power signals are processed using a 1D convolutional neural network in their original form. Then these signals are converted into images and processed using a 2D convolutional neural network. Then, feature vectors generated by 1D and 2D convolutional neural networks are combined. Finally, this combined vector is classified by a fully connected layer. The proposed method is well suited to the nature of signal processing. It is a novel approach that covers the steps of an expert examining a signal. The proposed framework is compared with other state-of-the-art power quality disturbance classification methods in the literature. While the proposed method's classification performance is relatively high compared to other methods, the computational complexity is almost the same.Öğe A novel hybrid model in the diagnosis and classification of Alzheimer's disease using EEG signals: Deep ensemble learning (DEL) approach(Elsevier Sci Ltd, 2024) Nour, Majid; Senturk, Umit; Polat, KemalRecent years have witnessed a surge of sophisticated computer-aided diagnosis techniques involving Artificial Intelligence (AI) to accurately diagnose and classify Alzheimer's disease (AD) and other forms of Dementia. Despite these advancements, there is still a lack of reliable and accurate methods for distinguishing between (AD) and Healthy Controls (HC) using Electroencephalography signals (EEG). The main challenge is finding the right features from the intricate spectral-temporal EEG data, which can provide information sufficient for diagnosis. This study proposes a new approach integrating Deep Ensemble Learning (DEL) and 2-dimensional Convolutional Neural Networks (2D-CNN) to address these issues. Combining state-of-the-art supervised deep learning algorithms within an ensemble model architecture aims to accurately diagnose and classify EEG signals of AD and HC subjects. Public EEG-based Alzheimer's datasets have been classified in the DEL model without applying any feature extraction after cleaning from noise and artifacts. Furthermore, the proposed DEL model used 5 different 2D-CNN models as internal classifiers. As a result, the EEG-based DEL model proposed for the first time provided high accuracy in AD classification. The proposed DEL model reached an average accuracy of 97.9% in AD classification due to 5 cross-fold training. In conclusion, this work renders that incorporating ensemble learning techniques into automotive health applications create extensible and stable AI models needed for computer-aided diagnostic. However, although the reported results and evaluation are promising, further efforts will need to be made to improve the accuracy of our proposed model. In addition, a fine-grid evaluation will be necessary to accurately understand potential impacts in clinical applications, such as earlier diagnosis or treatment decisions.Öğe A novel medical diagnosis model for covid-19 infection detection based on deep features and bayesian optimization(Elsevier, 2020) Nour, Majid; Cömert, Zafer; Polat, KemalA pneumonia of unknown causes, which was detected in Wuhan, China, and spread rapidly throughout the world, was declared as Coronavirus disease 2019 (COVID-19). Thousands of people have lost their lives to this disease. Its negative effects on public health are ongoing. In this study, an intelligence computer-aided model that can automatically detect positive COVID-19 cases is proposed to support daily clinical applications. The proposed model is based on the convolution neural network (CNN) architecture and can automatically reveal discriminative features on chest X-ray images through its convolution with rich filter families, abstraction, and weight-sharing characteristics. Contrary to the generally used transfer learning approach, the proposed deep CNN model was trained from scratch. Instead of the pre-trained CNNs, a novel serial network consisting of five convolution layers was designed. This CNN model was utilized as a deep feature extractor. The extracted deep discriminative features were used to feed the machine learning algorithms, which were k-nearest neighbor, support vector machine (SVM), and decision tree. The hyperparameters of the machine learning models were optimized using the Bayesian optimization algorithm. The experiments were conducted on a public COVID-19 radiology database. The database was divided into two parts as training and test sets with 70% and 30% rates, respectively. As a result, the most efficient results were ensured by the SVM classifier with an accuracy of 98.97%, a sensitivity of 89.39%, a specificity of 99.75%, and an F-score of 96.72%. Consequently, a cheap, fast, and reliable intelligence tool has been provided for COVID-19 infection detection. The developed model can be used to assist field specialists, physicians, and radiologists in the decision-making process. Thanks to the proposed tool, the misdiagnosis rates can be reduced, and the proposed model can be used as a retrospective evaluation tool to validate positive COVID-19 infection cases. (C) 2020 Elsevier B.V. All rights reserved.Öğe A novel tilt and acceleration measurement system based on hall-effect sensors using neural networks(Hindawi Ltd, 2022) Nour, Majid; Daldal, Nihat; Kahraman, Mehmet Fatih; Sindi, Hatem; Alhudhaif, Adi; Polat, KemalA tilt sensor is a device used to measure the tilt on many axes of a reference point. Tilt sensors measure the bending position according to gravity and are used in many applications. Slope sensors allow easy detection of direction or slope in the air. These tilt gauges have become increasingly popular and are being adapted for a growing number of high-end applications. As an example of practical application, the tilt sensor provides valuable information about an aircraft's vertical and horizontal tilt. This information also helps the pilot understand how to deal with obstacles during flight. In this paper, Hall-effect effective inclination and acceleration sensor design, which makes a real-time measurement, have been realized. 6 Hall-effect sensors with analog output (UGN-3503) have been used in the sensor structure. These sensors are placed in a machine, and the hall sensor outputs are continuously read according to the movement speed and direction of the sphere magnet placed in the assembly. Hall sensor outputs produce 0-5 Volt analog voltage according to the position of the magnet sphere to the sensor. It is clear that the sphere magnet moves according to the inclination of the mechanism when the mechanism is moved angularly, and the speed of movement from one point to the other changes according to the movement speed. Here, the sphere magnet moves between the hall sensors in the setup according to the ambient inclination and motion acceleration. Each sensor produces analog output values in the range of 0-5 V instantaneous according to the position of the spheroid. Generally defined, according to the sphere magnet position and movement speed, the data received from the hall sensors by the microcontroller have been sent to the computer or microcomputer unit as UART. In the next stage, the actual sensor has been removed. The angle and acceleration values have been continuously produced according to the mechanism's movement and output as UART. Thanks to the fact that the magnet is not left idle and is fixed with springs, problems such as vibration noises and wrong movements and the magnet leaning to the very edge and being out of position even at a slight inclination are prevented. In addition, the Hall-effect sensor outputs are given to an artificial neural network (ANN), and the slope and acceleration information is estimated in the ANN by training with the data obtained from the real-time slope and accelerometer sensor.
