Hamed Nassar

An efficient algorithm for incremental mining of temporal association rules

This paper presents the concept of temporal association rules in order to solve the problem of handling time series by including time expressions into association rules. Actually, temporal databases are continually appended or updated so that the discovered rules need to be updated. Re-running the temporal mining algorithm every time is ineffective since it neglects the previously discovered rules, and repeats the work done previously. Furthermore, existing incremental mining techniques cannot deal with temporal association rules. In this paper, an incremental algorithm to maintain the temporal association rules in a transaction database is proposed. The algorithm benefits from the results of earlier mining to derive the final mining output. The experimental results on both the synthetic and the real dataset illustrate a significant improvement over the conventional approach of mining the entire updated database.

Automated testing of java menu-based GUIs using XML visual editor

A common method for GUI testing is the capture and replay (CR) technique. While the commonly used of CR technique can not be used until an application under test (AUT) is completely developed. In this paper we propose a specification-driven approach to test menu-based Java GUI programs as an alternative to the CR technique. We introduce a GUI-event test specification language based on XML from which an automated test engine is generated. The popularity of XML as a scripting language is ease of use and its standard format. Along with our visual test scripts editor have made writing a test script using our proposed method is easy task. Beside the ability to test AUT before being completely developed, we have implemented another visual module that can generate the test specification file for an already existing AUT by using only the executable file of the program (.class). The engine generates GUI events and captures its responses automatically to verify the results of the test cases, which are written to a test log file. This engine supports the menu-based GUIs that became very popular these days in programs interfaces.

DARM: Decremental Association Rules Mining

Frequent item sets mining plays an important role in association rules mining. A variety of algorithms for finding frequent item sets in very large transaction databases have been developed. Although many techniques were proposed for maintenance of the discovered rules when new transactions are added, little work is done for maintaining the discovered rules when some transactions are deleted from the database. Updates are fundamental aspect of data management. In this paper, a decremental association rules mining algorithm is present for updating the discovered association rules when some transactions are removed from the original data set. Extensive experiments were conducted to evaluate the performance of the proposed algorithm. The results show that the proposed algorithm is efficient and outperforms other well-known algorithms.

Fault-tolerance for Switching Networks

Switching networks (SNs) are typically placed at the center of the systems they serve. Consequently, if an SN fails, system performance is bound to deteriorate or, in the worst case, the system may collapse totally. To limit these troubling consequences, the SN has to be fault tolerant [11]. Loosely defined, a fault tolerant SN is a SN that can still work in the presence of faults. Unfortunately, there is no universal, unique definition which can be applied to a network to see whether it is fault tolerant or not. To a great extent, fault tolerance is “in the eyes of the designer!” Specifically, a designer arbitrarily sets for the network in question a fault tolerance model [3, 4], defining what a fault is, how many such faults the network can (or should) tolerate, and exactly what is meant by “tolerate.”

An efficient technique for incremental updating of association rules

Recent developments in information technology resulted in rapid accumulation of huge amounts of data. Therefore, the task of managing large databases efficiently, in addition to quickly retrieving useful information and mining association rules, has become increasingly important. In this paper, an efficient technique is proposed for updating the discovered association rules in incremental databases. The experimental results show that the proposed algorithm is quite efficient and outperforms other well known algorithms.

Design and Analysis of a Channel Assignment Scheme for CDMA/TDMA Mobile Networks

In this paper, a channel assignment scheme is proposed for use in CDMA/TDMA mobile networks carrying voice and data traffic. In each cell, three types of calls are assumed to compete for access to the limited number of available channels by the cell: new voice calls, handoff voice calls, and data calls. The scheme uses the movable boundary concept in both the code and time domains in order to guarantee the quality of service (QoS) requirements of each type. A traditional Markov analysis method is employed to evaluate the performance of the proposed scheme. Measures, namely, the new call blocking probability, the handoff call forced termination probability, the data call loss probability, the expected number of handoff and the handoff link maintenance probability are obtained from the analysis. The numerical results, which are validated by simulation, indicate that the scheme helps meet the QoS requirements of the different call types.

Using hidden convexity in structured communication problems

Generalized minimum mean squared error (GMMSE) detector has a bit error rate (BER) performance that is almost the same as minimum mean squared error (MMSE) detector and it has the advantage that it does not require the knowledge of the noise power. However, the computational complexity of GMMSE detector is higher than that of MMSE detector. In this paper, the computational complexity of the GMMSE detector is reduced by using the hidden convexity combined with the idea of using the circular approximation of the banded Toeplitz channel matrix. In this case, the solutions are obtained by computing the EVD of the circular matrix using FFT/IFFT such that, the required EVD implies no additional effort.

Design and analysis of a TDMA call assignment scheme for cellular networks

In this paper, we introduce a TDMA channel assignment scheme suitable for personal communication services (PCS) networks. This scheme, called Handoff TDMA (H-TDMA), assumes three type of traffic: new voice calls, handoff voice calls and data packets, all competing for a finite number of channels. In the present scheme, priority is given to handoff voice calls over data calls, while a number of channels is exclusively dedicated to data calls. The performance of the scheme is analyzed using a Markov chain model. We evaluate major performance metrics such as forced termination probability, blocking probability, and data packets loss probability. The analytical results obtained are validated by simulation and good agreement is obtained.

Performance Analysis of Finite Buffered Multistage Interconnection Networks Routing Two-Class Traffic

In this article, we analyze the performance of a twoclass Multistage interconnection network with outputbufered SEs, using a discrete priority queueing model. We establish equations for steady state system occupancy distribution. These equations are then used to find important MIN measures, such as throughput, and loss probability. We also discuss the effect on these measures of the different operating parameters, such as the packet arrival rate and buffer size. We validate our results using simulation.

3D object compression using enhanced stripification technique

3D meshes are one of the most common representations of 3D objects that have been used in graphics applications. They often require a huge amount of data for storage and/or transmission. In this paper, we present a new compression algorithm based on stripification of the geometric models that enables us to progressively visualize the 3D models during their transmission. The proposed algorithm encodes the geometry and the connectivity of the input model in an interwoven fashion. The main idea is to store 3D objects as strips files. The algorithm achieves compression ratios above 61 : 1 over ASCII encoded formats resulting in faster transmission and rendering of complex graphical objects.