Thesis Title: An Approach for Finding a Best Path Routers
Author: Mohammad Hafez Abdul-Rahim Mustafa, Supervisor: Dr. Saad Al-Mahdawy, Year: 2008
Faculty: Information Technology, Department: Computer Science

Abstract: One of the prevailing tendencies of the modern stage of development of information technology is the telecommunication technologies integration based on computer networks which become more complex and the traffic load increases. There is a need to determine the routing traffic within a network so as to minimize the number of communication channels used. To reduce the risk of being unable to handle traffic required to find the best and optimal path from the source to the destination and to minimize the total cost of the system operation. This thesis will focus on the methodology that implements hybrid dynamic routing protocol that can solve congestion problem and hacking problem using Adaptive Genetic Algorithm. The new suggested structure combines different solutions to select the optimal path. Such structure will take into consideration different circumstances related to high load and utilization on advanced Wide Area Networks due to Security gaps and probable attacks and network activities.

Keywords: Routing, Traffic, Hybrid Dynamic Routing Protocol, Adaptive Genetic Algorithm, Wide Area Networks

Thesis Title: Speaker Recognition Using Neural Network Model
Author: Ayham Rasem Fayz Jaroun, Supervisor: Dr. Venus W. Samawi, Year: 2008
Faculty: Information Technology, Department: Computer Science

Abstract: The use of biometric information has been known widely for both person identification and security application. It is common knowledge that each person can be identified by the unique characteristics of one or more of his / her biometrics. One of the biometrics that a person can be identified by is the unique characteristics of his/her voice. This work is concerned with the study of using voice as biometric information (i.e. Speaker Recognition) for controlling access to the facility that need to be protected from the intrusion of unauthorized persons. The main significance of this work is to study a number of experimental investigations of using neural networks to recognize speakers and suggest the best neural network architecture leading towards the goal of higher accuracy for speaker recognition. Therewithal, attempt to draw a conclusion about the recommended feature-set (based on Wavelet Transformation) that could be used to discriminate speakers and improve the overall performance. To do this, features are extracted from different levels of continues wavelet transformation (three maximum values with their locations and level number in addition to mean and standard division of each level), these features are used to train three of Speaker Recognition Neural Networks (Adaptive Neural Network, Feed-forward Backpropagation Neural Network, and Learning Vector Quantization Neural Network). Experimental results that show the classification accuracy of each classifier were introduced. Then these results were analyzed and compared to arrive at the best neural network (as speaker recognizer), and decide which of the feature (or combination of them) among the above set leads to the minimum classification error rate (i.e. has the best discrimination ability).

Keywords: Biometric Information, Speaker Recognition, Wavelet Transformation, Adaptive Neural Network, Feed-forward Backpropagation Neural Network, and Learning Vector Quantization Neural Network

Thesis Title: Using Aspect-Oriented Programming to Secure the Broken Authentication and Session Management On Web Application
Author: Sandrella Ibrahim Kamel Mahjoub, Supervisor: Prof. Said Ghoul, Year: 2008
Faculty: Information Technology, Department: Computer Science

Abstract: Web applications are becoming more and more popular every day. Many web applications made life easier. We have webmail, online retail sales, online bills payment, flights check-in and status, wikis, multiplayer online role-playing games, and many others. Due to the increase of web application usage the security become as a most critical aspects which no one can use or trust such application without guarantee security. Web application concerns contains business concern which takes the principal view, and other concerns as security, logging, tracing, performance ,this all concerns take lower level of importance in the design of a web application. The Aspect-Oriented Programming (AOP) paradigm focuses on the identification, specification and representation of crosscutting concerns and their modularization into separate functional units given more than one concern higher level of importance. In this thesis, we applied the Aspect-Oriented Programming approach to enhance the security of web applications. That by solving a security problem, the problem is the interception of the client and server connection by a third party to stolen session Id after success login and use the stolen session Id to a legal access to the server response for illegal user.

Keywords: Web applications, Aspect-Oriented Programming

Thesis Title: Using Artificial Neural Networks Models for Predicting Pancreas Behavior
Author: Bassam Abdul-Rahman Abed, Supervisor: Dr. Rashid Al-Zubaidy, Year: 2008
Faculty: Information Technology, Department: Computer Science

Abstract: Artificial Neural Network (ANN) is currently a 'hot' research area in medicine and this work is based on predicting the behavior of an organ of a human body called pancreas by using neural networks. Neural networks, with their remarkable ability to derive meaning from complicated or imprecise data, can be used to extract patterns and detect trends that are too complex to be noticed by either humans or other computer techniques. Their ability to learn by example makes them very flexible and powerful. So we have data set from 102 of three different groups of subjects. From data set examples, neural networks were learned by using two algorithms, Radial Basis Function (RBF) and General Regression Neural Network (GRNN). After learning, a simulation (testing) for learned data has been made and a comparison between both algorithms has been done subject to the learning performance. Furthermore a comparison between RBF and GRNN algorithms was done based on the ability of both networks to generalize input data not seen before.

Keywords: ANN, GRNN, RBF

Thesis Title: A Data Mining Approach for Categorizing Web Documents
Author: Kifaya Said Qaddoum, Supervisor: Dr. Fadi Fayez Thabtah, Year: 2008
Faculty: Information Technology, Department: Computer Science

Abstract: The core step for KDD is Data Mining. Data Mining applies efficient algorithms to extract interesting patterns and regularities from the data. As volume of information in digital from increases, the use of Text Categorization techniques, which aim at finding relevant information, becomes more necessary. To improve the quality of the classification process from textual data sets, Associative Classification, which utilizes the association rule discovery techniques to construct classification systems, is evaluated in this thesis. Particularly, we developed an associative classification vertical mining algorithm representation in order to improve the accuracy of the classification phase, and to reduce the size of the memory required to store intermediate TIDS in the mining process. Considering the fact that vertical data structure supports fast frequency counting via intersection operations on transaction identifiers (TIDS), thos should improve accuracy and decrease memory usage. This thesis demonstrates the problem of using Associative Classification to solve Text Categorization problem, and utilize Diffset structure as a mining approach.

Keywords: Data Mining,Associative Classification, Transaction Identifiers (TIDS), Text Categorization, Diffset