Journal of Al-Qadisiyah for Computer Science and Mathematics

In this article we conceder the logistic regression model with high leverage points. For the logistic regression model with a binary response, we suggested a new robust approach called robust logistic regression (RLR) based on the robust mahalanobis distance (RMD) which depends on the minimum volume ellipsoid (MVE) estimators. The RMD is computed by using the algorithm of stochastic gradient descent (SGD). In order to assist the new suggested approach we compare it with some existing method such as maximum likelihood estimator and robust M-estimator in logistic regression model. The simulation study points that the RLR has supreme performances throw some measurement comparison.

In recent year, the use of the internet and cloud-based application has increased tremendously. With this lot of internet apps and social network have been developed. This apps and networks are growing very rapidly. Data on this network belongs to different data sizes and structures. This data in simple language is called big data. With the increasing demand for services, all the big data services are now automated, which stores and retrieve consumer's data. This data is in a different structure which makes it difficult to handle. This can include an example of road traffic data, different vehicle or person or location or atmosphere can produce a different set of data. Sometimes this data is structured while some times it may contain repeated, null or noisy data. This data may contain images and videos. The traditional database cannot handle this type of unstructured data. The best to solve this dimensionality problem is by using a dimensionality reduction technique. This paper provides the concept of big data collection, its analysis, storage and handling issues, security challenges, and other data handling technique.

Automatic speaker recognition systems use the machines to recognize an individual via a spoken sentence. Those systems recognize a specific individual or confirm an individual’s claimed identity. The most common type of voice biometrics is the Speaker Recognition. Its task focused on validation of a person’s claimed identity, using features that have been obtained via their voices. Throughout the last decades a wide range of new advances in the speaker recognition area have been accomplished, but there are still many problems that need solving or require enhanced solutions. In this paper, a brief overview of speech processing is given firstly, then some feature extraction and classifier techniques are described, also a comparative and analysis of some previous research are studied in depth, all this work leads to determine the best methods for speaker recognition. Adaptive MFCC and Deep Learning methods are determined to be more efficient and accurate than other methods in speaker recognition, thus these methods are recommend to be more suitable for practical applications.

Recently, deep learning models based on convolutional neural networks (CNN) have been used in image steganalysis problems. In this paper, we present different architecture of CNN with dual tree complex wavelet transform for preprocessing before input images put into system. The main task of this transform is for exploiting the difference between cover and stego images through shift variance property. The net consists of five successive convolutions layers. Each one following by normalization and pooling layers ends with fully connected layer. The performance of system is evaluated through accuracy, precision, recall and f-score measures. The results show effectiveness of it with more than 0.9 precision values. HUGO, WOW and UNIWARD algorithms selected for implementation.

In life, a chaotic system has many applications in different fields, including physics, biology, communication, and cryptography. In this study, a new hyperchaotic system is introduced. This hyperchaotic system is a two-dimensional system that is based on three maps-namely, logistic, iterative chaotic, and Henon maps. The dynamics of this system are investigated using maximal Lyapunov exponents, bifurcation diagrams, phase portraits, basin of attraction, and complexity via entropy. This system shows highly complicated dynamics. On the basis of the proposed system, a new algorithm for image encryption is also introduced. Confusion and diffusion can be achieved with this algorithm, which are fundamental demands. The stochastic behavior of this system is used to reinforce the security of the encrypted image. The image is divided into four parts, each of which uses a different random key established by the proposed chaotic system. The security of this cryptosystem is validated on the basis of key security parameters and common attacks.

A comparative study was conducted in this paper , between three algorithms which (Harris , Shi-Tomasi , FAST ) interested-points detection to identified the features that required to match , recognize and track objects in images noisy .
Detect the interested-points in image noisy one of the most challenges in field of image processing . The noise consider the main cause for damage the natural images during the acquisition and transition , and detect the interested-points of these images doesn’t give the desired results , so eliminating noise from this images is very important , Non-local means approach is applied for solve this problem .

Spatial indexes, like the ones that are based on Quad Tree, are important in spatial data-bases for the effective implementation of queries with spatial constraints, particularly in the case where queries include spatial links. The quad trees are a very interesting subject, given the fact that they give the ability to solve problems in a way that focuses only on the important areas with the highest density of information. But it is not without the disadvantages because the search process in the quartile suffers from the problem of repetition when reaching the terminal node and return to the behaviour of another way in the search and lead to the absorption of large amounts of time and storage. A database management system can handle data very easily if the object is one-dimension (sequential).
In this paper, improve the quad tree by combining one of the space filling curve types, including the Hilbert curve and the Z-ordering curve with a quad tree. It will convert from two-dimensional to one-dimensional and sequentially search and end the problem of repetition whenever it reaches a terminal node Ordinary quad tree. Resulting in reduced storage space requirements and improved implementation time.

In this paper, the commuting and centralizing of symmetric reverse ∗-n-derivation on Lie ideal are studied and the commutativity of prime ∗-ring with the concept of symmetric reverse ∗-n-derivations are proved under certain conditions.

The main object of the present paper is to, introduce the. class of meromorphic univalent functions Involving! hypergeomatrc function .We obtain~ some interesting geometric properties according to coefficient inequality , growth and distortion theorems , radii of starlikeness and convexity for the" functions in our subclass.

The aims of this study are to introduce the acceleration methods which are called inverse triangular acceleration methods and inverse hyperbolic triangular acceleration methods, which we generally call Al-Tememe's acceleration methods of the second kind discovered by (Ali Hassan Mohammed). It is useful to improve the numerical results of continuous integrals in which the error is of the 4th order, and regarding to accuracy, the number of used partial intervals and how fast is to get results especially to accelerate the results that can be got by using Simpson's method. Also, it is possible to utilize it in improving the numerical results of differential equations ,where the main error is of the forth order.