Abstract This paper presents an e-Learning tool for mastering the back-propagation neural network architecture. A short review of the existing tools is presented. It is developed using MS Visual C++. The tool’s functionality can be summarized as: First, at its highest-level, it operates two basic modes: the training mode and the recall mode. Second, while it is in training, it has... This paper presents an e-Learning tool for mastering the back-propagation neural network architecture. A short review of the existing tools is presented. It is developed using MS Visual C++. The tool’s functionality can be summarized as: First, at its highest-level, it operates two basic modes: the training mode and the recall mode. Second, while it is in training, it has two sub-modes: the learning-mode and the application-mode. In learning mode, the software generates textoutput traces corresponding to the top-down design steps of the NN-architecture. The generated numeric traces have dualusage, either they can be used learning purposes or for generating class room tests. While in application-training mode, the tool displays only the input-output relations – the values before and after the training. In this mode the tool also generates a cumulative error-index to monitor the progress of the network training. Third, it enables the user to enter the network training termination criteria. Fourth, at the end of the network training, it is stores the trained network into a text-file. Fifth, in the test or recall mode, the trained network is retrieved from a stored-file, it then generates the network response corresponding to the entered test input. The e-Learning tool is tailored for mastering, class room teaching, and test generation of the BP-NN-architecture. Read More Read Less

Abstract One of the most important fields in image processing is the image inpainting or image retouching with the aim to restore damaged images or remove a selected object from an image. Many recent works in the image inpainting focus on combining methods in order to obtain more accurate results. In this paper we proposed a new algorithm that combines Kmeans... One of the most important fields in image processing is the image inpainting or image retouching with the aim to restore damaged images or remove a selected object from an image. Many recent works in the image inpainting focus on combining methods in order to obtain more accurate results. In this paper we proposed a new algorithm that combines Kmeans algorithm and the partial differential equation (PDE). Read More Read Less

Abstract Mutation Testing is a fault–based software testing technique that has been neglected by industry for a long time because of its high cost. The rising research in this field has resulted in the development of several approaches that aim to reduce the cost of testing and to assess the quality of mutants’ generation and destruction. Some approaches suggest decreasing the... Mutation Testing is a fault–based software testing technique that has been neglected by industry for a long time because of its high cost. The rising research in this field has resulted in the development of several approaches that aim to reduce the cost of testing and to assess the quality of mutants’ generation and destruction. Some approaches suggest decreasing the number of operators, while others intend to reduce the execution time. The goal of this paper is to evaluate the effectiveness of 10-selective mutation approaches. The experimental part of this research was based on testing seven Java programs using MuClipse tool. The analysis of the findings has proved that the cost of mutation testing can be minimized by selecting a subset (10 operators) from mutation operators. Read More Read Less

Abstract Modeling and performance analysis of the media access control layer are extremely significant. Therefore, the performance evaluation helps a researcher in discovering the inherent cause of many problems, and may even suggest possible solutions. This is why much of the research work aimed to achieve an accurate evaluation. However, a lot of work does not consider the main difference between... Modeling and performance analysis of the media access control layer are extremely significant. Therefore, the performance evaluation helps a researcher in discovering the inherent cause of many problems, and may even suggest possible solutions. This is why much of the research work aimed to achieve an accurate evaluation. However, a lot of work does not consider the main difference between the busy probability and the collision probability in analytical mathematical models. Some of them consider both probabilities to be the same or ignore the busy probability and consider only the collision probability, which is not accurate. This paper proposes a new performance evaluation by creating a new mathematical model to compute a packet transmission probability for IEEE 802.11 DCF. Therefore, we enhance Bianchi’s model by adding a new state to represent the busy probability in the mathematical model, which helps to achieve the highest accuracy performance evaluation of IEEE 802.11 DCF. Read More Read Less

Abstract We report on fractional modeling and identification of non-integer systems by least squares method (LS). A new approach to identify fractional differentials equations (FDE) is proposed. Such a technique presents a linear model to estimate system parameters, as well as non-integer orders from temporal data (Hn model). The identified parameters provide original solution to the Output Error method initialization (OE),... We report on fractional modeling and identification of non-integer systems by least squares method (LS). A new approach to identify fractional differentials equations (FDE) is proposed. Such a technique presents a linear model to estimate system parameters, as well as non-integer orders from temporal data (Hn model). The identified parameters provide original solution to the Output Error method initialization (OE), and demonstrate validity and effectiveness of the proposed approach. Read More Read Less