It is very important to analyze the logs on the Firewall devices and control the internet traffic according to these analysis results. In this study, some logs obtained with the Firewall Device used at Firat University are classified using multiclass support vector machine (SVM) classifier. Linear, polynomial, sigmoid and Radial Basis Function (RBF) functions are used as the activation function for SVM classification. In order to measure the performance of the classifier, the comparison was made by finding the measurement values of sensitivity, recall and their harmonic mean F1 Score. In this study, 65532 instances have been examined using 11 features. The feature that characterizes any personal data in the selected data has not been used. The Action attribute is selected as the class from these attributes. The “allow”, “deny”, “drop” and “reset-both” parameters have been implemented for the Action class. Activation functions have been tried and the SVM responses have been evaluated so as to obtain the maximum recall and precision values in the SVM classifier. It was tried to obtain the best activation function for F ₁ score value. Receiver Operating Characteristic (ROC) curves were also created for each of the classes. At the end of the study, the activation functions from which the desired SVM responses are obtained are given by comparison.

Estimation of dimension parameters for an electrical machine has great importance before manufacturing. For this reason, analytical design should be performed in an optimum form. While motor analysis is accomplished by package programs, initial size parameters are intutivily provided and then various trials are examined to get optimum results. In this study, we are trying to find dimensional and electrical parameters generating mathematical equations in analytic approaches for In-Wheel Switched Reluctance Motor (IW-SRM), which will be employed by Electric Vehicle (EV). Therefore, optimum motor parameters for required speed and torque have been estimated by solving generated equations for in-wheel SRM with 18/12 poles via MATLAB. Using the parameters, analysis of in-wheel SRM has been carried out 3D Finite Element Method (FEM) by Ansoft Maxwell 15.0 Package Software. Consequently, the accuracy of the estimated parameters has been validated by the results of Maxwell 3D FEM.

Recent advances in object detection and classification has accelerated by virtue of Convolutional Neural Networks (CNN) and large-scale image datasets like ImageNet. In this study, we aim to classify objects such as knives and weapons on the image data contained in electronic evidence, which may be criminal objects in judicial cases. To this end, we make this process autonomous by contributing to the examination of electronic evidence via artificial intelligence. For training of the CNN model presented in this paper, we use 10000 training data handled from ImageNet. The model with best parameter settings has been achieved 94.37% classification accuracy on 2500 test samples. We also investigate the effect of different parameters used in the network to increase the classification accuracy and try to achieve a set of optimal hyper parameters by tuning best possible parameters. As a result of the work, we present a detailed comparative study on binary object classification.