The spine is composed of pieces of bone called vertebrae that lie between the skull and the tailbone. Various medical conditions can affect the spine. In this study, two types of degenerative diseases, scoliosis, and spondylolisthesis, were studied. Deep AI architectures have recently enabled further disease diagnosis innovation using medical images. Various traditional and deep learning studies use medical images for disease diagnosis in the literature. This study aims to classify spine X-ray images according to three possible conditions (Normal, Scoliosis, and Spondylolisthesis) and to exploit the potential of these X-ray images to detect possible diseases occurring in the spine. The performance of deep learning models and optimization algorithms used in this process was evaluated. The study uses a data set created and/or analyzed during an existing study. This data set consists of images that belong to three different classes: scoliosis, spondylolisthesis, or x-ray images of normal (i.e. healthy) individuals. A total of 338 spine X-ray images, 188 scoliosis images, 79 spondylolisthesis images, and 71 normal images. Six different deep-learning architectures have been used in the study. These architectures are Alexnet, GoogLeNet, ResNet-18, ResNet-50, ResNet-101, and EfficientNet-bo. While working on these deep architectures, each model has been evaluated using different optimization algorithms. These optimization algorithms are RmsProp, SGDM, and Adam. According to the classification processes, the deep learning model with the highest accuracy value was Alexnet, and the optimization algorithm used with it, Sgdm (99.01%), and the training time lasted 38 seconds. According to the classification processes, the deep learning model with the fastest completion time (30 seconds) was Alexnet and the optimization algorithm used with it was RmsProp. An accuracy rate of 98.02% has been obtained in the training of this model.

Su dağıtım şebekelerinde, borulardaki çatlaklar ve arızalardan dolayı ciddi miktarda su kaybı yaşanmaktadır. Bu arızaların kısa zamanda tespit edilerek onarılması, su ve buna bağlı oluşan gelir kaybının önlenmesi için oldukça önemlidir. Dağıtım şebekelerinde yüksek maliyete sahip genel onarım işlemleri yerine, arızanın kesin konumunun bulunup sadece o bölgede çalışma yapılması onarım maliyetlerini azaltacaktır. Her ne kadar yüzeysel boru dinleme cihazları bu ihtiyaca bir çözüm olarak görünse de, dış ortam seslerinden etkilenmesi bu yöntemin verimliliğini düşürdüğünden tercih edilmemesine neden olmaktadır. Ticari olarak piyasada mevcut olan modern GPS temelli sistemler, büyük çaplı su borularında çalışabilir (≥6 inch) ve yüksek maliyetlere sahiptir. Bu çalışmada daha küçük çaplı borularda ve GPS sistemine ihtiyaç duymadan çalışabilecek bir sızıntı tespit robotu prototipinin ön çalışması sunulmuştur. Bu ön çalışmada, boru içerisinde suyun itme kuvvetiyle hareket edecek robotun tasarımı, üretimi, konum ve sızıntı tahmin yazılımları gerçekleştirilmiştir. Konum tahmini, 9-dof IMU sensör (3D-ivme, 3D-jiroskop ve 3d-manyetometre) verilerinin Genişletilmiş Kalman Filtresi içerisinde kullanımıyla yapılmaktadır. Sızıntı tahmini, anlık kaydedilen ses verisindeki tepe noktalara karşılık gelen konumun tespitini içermektedir. Yapılan deneysel çalışmalarda, toplam 118m gezintide sonucunda sızıntı konumu tahmin hatasının yaklaşık 0,25m olduğu görülmüştür.

In this study; time series electromyography (EMG) data have been classified according to hand movements using wavelet analysis and deep learning. A pre-trained deep CNN (Convolitonal Neural Network-GoogLeNet) has been used in the classification process performed with signal processing, by this way the results can be obtained by continuous wavelet transform and classification methods. The dataset used has been taken from the Machine Learning Repository at the University of California. In the data set; EMG data of 5 healthy individuals, 2 males and 3 females, of the same age (~20-22 years) are available. Data; It consists of grasping spherical objects (Spher), grasping small objects with fingertips (Tip), grasping objects with palms (Palm), grasping thin/flat objects (Lat), grasping cylindrical objects (Cyl) and holding heavy objects (Hook). It is desired to perform 6 hand movements at the same time. While these movements are necessary, speed and power depend on one's will. People perform each movement for 6 seconds and repeat each movement (action) 30 times. The CWT (Continuous Wavelet Transform) method was used to transform the signal into an image. The scalogram image of the signal was created using the CWT method and the generated images were collected in a data set folder. The collected scalogram images have been classified using GoogLeNet, a deep learning network model. With GoogLeNet, results with 97.22% and 88.89% accuracy rates were obtained by classifying the scalogram images of the signals received separately from channel 1 and channel 2 in the data set. The applied model can be used to classify EMG signals in EMG data with high success rate. In this study, 80% of data was used for educational purposes and 20% for validation purposes. In the study, the results of the classification processes have been evaluated separately for first and second channel data.

In this digitalized world, users of various software systems would like to securely make use of it at every stage from data generation to analysis. However, blocking these services by malicious people is also an undesirable phenomenon in our world. Since Distributed Denial of Service (DDoS) attack detection is important due to its increasing prevalence, this paper presents machine learning and hybrid approaches for DDoS detection. This study was performed on the popular CICIDS2017 and CIC-DDoS2019 datasets used in DDoS attack detection. Also, an alternative hybrid dataset is created by combining these two datasets. This study initially employed Decision Trees (DT), Random Forest (RF), K-Nearest Neighbors (KNN), and Support Vector Machines (SVM) machine learning algorithms on the specified datasets, thereafter conducting a comprehensive assessment of each model's efficacy. We further evaluated the datasets employing hybrid modeling that integrates two machine learning methods to enhance performance, accuracy, and dependability by leveraging their respective strengths. The investigation demonstrated that hybrid models may get an accuracy of up to 99.91% on complex data sets. In our research, we combined two important datasets to construct an alternative to those utilized in existing literature. The hybrid application of machine learning methods markedly enhanced DDoS detection accuracy and optimized performance on complex datasets relative to hybrid versions of established approaches. Moreover, our results aim to improve the efficiency and flexibility of cybersecurity detection techniques and to create a foundation for future research.

Breast cancer is a life-threatening cancer worldwide and is commonly seen in women. Early and accurate diagnosis is a key to successful treatment and better survival rates. In this study, using 569 samples from the Wisconsin Diagnostic Breast Cancer dataset, the classification performances of Support Vector Machines (SVM), Adaptive Boosting (AdaBoost), Random Forest (RF), Naïve Bayes (NB) and k-Nearest Neighbor (kNN) algorithms are compared. To improve accuracy, the train dataset was expanded from 381 to 1524 samples using a three-stage method involving random scaling, Gaussian noise, and mathematical transformations. Test dataset remained unchanged. Statistical analyses confirmed that the augmented data presented sufficient variability to enhance generalization while preserving the original distribution. Model performance analysis was conducted using accuracy, precision, sensitivity, F1 score, FDR, MCC and AUC values. Compared to the original dataset, the accuracy of the SVM algorithm improved from 98.94% to 99.47% with the data augmentation techniques. This result shows that data augmentation techniques are effective in improving the classification performance of machine learning models for breast cancer diagnosis.

Water is vital for all living beings, especially for a human. Automatic position detection of water leakage in water pipelines is very important to minimize the loss of labour, time, money spent on exploration and excavation in pipe inspection procedures. The main goal of detection is to prevent water loss. In this study, sensitive position detection, crack frequency band detection and external sphere studies of an in-pipe robot prototype have performed. During the precise position estimation, classical EKF, stationary region detection and location estimation using EHDE are performed with two different ANNs. In this way, online precise position estimation can be done on hardware that has not sufficient computational power for indoor robotic studies. In addition, the sound characteristics resulting from the crack at different hole size and water pressure intensity levels have investigated. Finally, a new sealing sphere design has devised and three different hydrophone sensor data have recorded on the SD card simultaneously. It has been found that the proposed ANN method has the performance to work online and can make a similar position estimation with the classical IMU position estimation method by 99%.