A three-stage learning algorithm for deep multilayer perceptron (DMLP) with effective weight initialisation based on sparse auto-encoder is proposed in this paper, which aims to overcome difficulties in training deep neural networks with limited training data in high-dimensional feature space. At the first stage, unsupervised learning is adopted using sparse auto-encoder to obtain the initial weights of the feature extraction layers of the DMLP. At the second stage, error back-propagation is used to train the DMLP by fixing the weights obtained at the first stage for its feature extraction layers. At the third stage, all the weights of the DMLP obtained at the second stage are refined by error back-propagation. Network structures an

The Strategy of employees' empowerment, organizational performance and total quality management represent three informative which were and still the hub of many studies and researches including the current study. The higher education is considered one of the most important services sectors and the main development as well as the civil progress in the society. It is the main source that provides the organizations with the human resources having scientific or practical professions. So Higher Education has been chosen especially Baghdad University as a sample of the study. The study problem is concentrated to find answers of the following   questions: What the level of the powers mandate in the organization concerning the

Background: Giant middle cerebral artery (MCA) aneurysms are surgically challenging lesions. Because of the complexity and variability of these aneurysms, a customized surgical technique is often needed for each case. In this article, we present a modified clip reconstruction technique of a ruptured complex giant partially thrombosed middle cerebral artery aneurysm.

Case description: The aneurysm was exposed using the pterional approach. Following proximal control, the aneurysm sac was decompressed. Then, we applied permanent clips to reconstruct the aneurysm neck. The configuration of the aneurysm mandated a tailored clipping pattern to account for resi

In the image processing’s field and computer vision it’s important to represent the image by its information. Image information comes from the image’s features that extracted from it using feature detection/extraction techniques and features description. Features in computer vision define informative data. For human eye its perfect to extract information from raw image, but computer cannot recognize image information. This is why various feature extraction techniques have been presented and progressed rapidly. This paper presents a general overview of the feature extraction categories for image.

The effect of the tensor term in the Skyrme interaction has been estimated in calculating the static and dynamic nuclear properties in sd and fp-shell model spaces nuclei. The nuclear shell gaps have been studied with different Skyrme parameterizations; Skxta and Skxtb with tensor interaction, SkX, SkM, and SLy4 without tensor interaction, and Skxcsb with consideration of the effect of charge symmetry breaking. We have examined the stability of N = 28 for 42Si and 48Ca. The results showed that the disappearance of the magicity occurs in the shell closure of 42Si. Furthermore, excitation energy, quadrupole deformation, neutron separation energy, pairing energy, and density profile have also been calculated. Quadrupole deformation indicates a

The traditional centralized network management approach presents severe efficiency and scalability limitations in large scale networks. The process of data collection and analysis typically involves huge transfers of management data to the manager which cause considerable network throughput and bottlenecks at the manager side. All these problems processed using the Agent technology as a solution to distribute the management functionality over the network elements. The proposed system consists of the server agent that is working together with clients agents to monitor the logging (off, on) of the clients computers and which user is working on it. file system watcher mechanism is used to indicate any change in files. The results were presente

The present research aimed to test the imagination of children, and may build sample consisted of (400) a baby and child, selected by random way of four Directorates (first Resafe, second Resafe ,first alkarkh , second alkarkh), in order to achieve the objective of research the tow researchers have a test of imagination and extract the virtual and honesty plants distinguish paragraphs and paragraphs and difficulty factor became the test consists of (32), statistical methods were used (Pearson correlation coefficient, coefficient of difficult passages, highlight paragraphs, correlation equation, an equation wrong Standard) the tow researchers have a number of recommendations and proposals.

This research deals with a very important subject as it tries to change the theoretical and scientific heritage and some professional rules adopted in the newsroom. Most media students have difficulties in writing news for press correctly. The researcher tries to identify the compatibility of what is published in local news agencies with professional and academic standards.
The research finds detailed editorial rules for a number of news formats which will play an important role in writing news for press easily, especially for the beginners and newcomers. Also, it discovers a new fact denying the beliefs of some researchers and writers in not having news conclusion in news edited according to the inverted pyramid pattern.
The re