Shengfei Shi Jianzhong Li Chaokun Wang

Shengfei Shi Jianzhong Li Chaokun Wang
School of Computer Science and Technology
Harbin Institute of Technology
Harbin Heilongjiang Province China
shengfei@hit.edu.cn

Abstract

With the development of computer networks and wireless communication technologies, more interests have been taken in mobile wireless ad hoc networks (MANET) that are only constructed by wireless hosts. In MANET database system, research must take into consideration issues both on clients and servers which are all mobile and power limited. Wireless sensor networks can be seen as data source in many applications based on MANET, such as the command system in battlefields and the large-scale geography monitor system. In these applications, mobile clients query the data generated by wireless sensor networks through mobile servers. Up to now, few papers have discussed the methods of data management in such complicated systems which include mobile clients, mobile servers and wireless sensor networks. The data of sensor networks can be treated as time series, and in most cases these time-series can construct a probability model such as AR, ARMA. Moreover, the communication cost can be saved and the future data can be predicted beforehand by using the data models.
In this paper, a novel and complicated system architecture is proposed. In order to deal with the data management problem, In-network modeling strategy, based on time-series analysis theory, is also proposed, which can significantly reduce the workload of communication and improve the performance of the system. By using in-network modeling methods, the traditional cache technology is endowed with new abilities, such as computable capability and stronger semantic representation faculty. The problem of mobile server’s mobility and the frequent disconnection of mobile clients are resolved by using TS-Cache. The experiments results of the simulation confirm the good performance of our algorithms under different situations.

Ramzi A. Haraty and Samer A. Khatib
Lebanese American University
Beirut, Lebanon
Email: rharaty@lau.edu.lb; khatibsam@hotmail.com

Abstract

TStemming has a large effect on Arabic information indexing and retrieval, at least partially due to the highly inflected nature of the language. Our work demonstrates the process of improving other stemmers, mainly that of [1]. We reached a recall difference of 28% over the work of [1]. The main part of improvement was due to the addition of more grammatical rules that facilitate the process of stemming.

Following this part, we implemented a procedure that extracts the temporal references from the texts. This procedure is highly dependable on the stemming process. A list of all the temporal references is used. The type of the temporal word decides the procedure to treat this word and gives the importance of this temporal reference. These conditions, with the help of the stemmer, produced an excellent result of 95% precision rate and of 91% recall rate.

Key words:  Indexer, stemmer, and temporal references.

DISTRIBUTED CONSTRUCTION OF WEAKLY CONNECTED DOMINATING SETS FOR CLUSTERING MOBILE AD HOC NETWORKS

FREDRICK JAPHET MTENZI
School of Computing
Dublin Institute of technology
Kevin Street, Dublin 8
Ireland
Email: Fred.Mtenzi@comp.dit.ie

Weakly connected dominating set (WCDS) has been proposed to cluster mobile ad hoc networks and be used as a virtual backbone. There have been some distributed approximation algorithms proposed in the literature for minimum WCDS. But none of them have constant approximation factors. Thus these algorithms can not guarantee to generate a WCDS of small size. Their message complexities may also be as large as . In this paper, we design a new distributed algorithm that outperforms the existing algorithms. This algorithm has an approximation factor of at most 5 and linear message complexity. Our algorithm requires only single-hop neighborhood knowledge and a message length of . So it is practical.

Key words:  Mobile ad hoc networks, Dominating Sets, Weakly Connected Dominating Set, virtual backbone.

K. Chandra Sekharaiah
Distributed Object Systems Group
Department of Computer Science & Engineering
JNTU College of Engineering
Hyderabad 500 072. India
Email: shakes123@sify.com
 

D. Janaki Ram
FDistributed Object Systems Lab
Department of Computer Science & Engineering
Indian Institute of Technology
Madras 600 036. India
Email: djram@lotus.iitm.ernet.in

Mohd. Abdul Muqsit Khan
Department of Computer Science
Moulana Azad National Urdu University
Hyderabad 500 032. India
abdul_muqsit_khan@yahoo.co.in

Abstract

Concerns are at the core of software engineering and composition. Concerns apply in terms of objects, methods, subjects, aspects, roles. This paper explores such various concerns from a comparison perspective. It concludes that none of the software composition techniques explored so far give adequate treatment in addressing the object schizophrenia problem for a complete solution. Our work is related to HRI and robot motion controls based on software composition.

Key words:  Software Composition Techniques - Roles, Subjects and Aspects, Object Schizophrenia Problem (OSP), Human Computer Interaction (HCI)

Semantic Web Service discovery and selection are a very time and resources consuming task. They require reasoning support for the matchmaking of the capabilities of services against user defined goals and constituent sub-goals, and for the mediation of the domain knowledge used to describe the different relevant aspects of services. This paper presents a performance study around the number of times the reasoner has to be used in nowadays initiatives. Such study lays the basis for an innovative approach inspired in the popular search engine Google, which tries to improve the performance of the whole process. The main idea is to carry the reasoning as an off-line task, storing the output for later reuse. It also elaborates on the idea of making service descriptions and goals available independently of registries or repositories, i.e. Web pages. Such idea permits to profit, extend and further reuse, well established concepts developed by popular search engines, thus assimilating service discovery and selection to any other type of search engine task.
 

Key words:  Semantic Web Services, WSMO, Google, Search engines.
 

OLA EL-SONOSY
Faculty of Computer and Information
Systems
Ain Shams University, Cairo, Egypt
Email: Sonoso@hotmail.com

AHMED HAMAD
Faculty of Computer and Information
Systems
Ain Shams University, Cairo, Egypt
Email: Amhamd13@yahoo.com

AMAL ELNAHAS
Faculty of Media Engineering and
Technology
The German University in Cairo,
Egypt
Email: amal.elnahas@guc.edu.eg

Abstract

Locating users as they move from one place to another in a cellular network is a key issue that allows unrestricted mobility, yet poses several challenging constraints to the network designers. In this paper, an enhanced adaptive location update scheme is proposed to decrease the total cost of the location management process. The proposed scheme relies on the deployment of a direction based location update scheme along with a simple prediction line paging technique to decrease the paging cost. The proposed protocol is implemented over both random walk and random waypoint mobility pattern. Results obtained proved a reduction in the overall cost up to 47% compared to the direction based location update scheme without prediction. Further, the accuracy of the prediction technique for users with varying speed is increased by issuing a location update message periodically. The slight increase in the update cost is compensated by the savings in the paging cost. This enhancement is implemented over two set of experiments with different cost coefficients. Both produced a reduction in the location management overall cost up to 26% compared to the proposed protocol without the enhancement.

Key words:  Location management, prediction technique, location updates, line paging, cellular networks

 

Focusing only on a knowledge delivery problem in distant learning systems, we can find course material selection with relation to an education level of particular student as a main shortcoming. The other equally week point to the mention above is immense burden for the course administrators, when number of students exceeds a few dozen or so. Then the number of people involved in planning, control, scheduling of classes and students’ progress assessment, increases in proportion to a number of students. The remedy for the presented above distant learning inconveniences and a way to improve efficiency of knowledge acquire process could be application of intelligent multi-agent system. In the paper, beyond theoretical consideration of multi-agent usefulness, model of a real multi-agent system (in couple variants) based on agents along with performance comparison will be presented.

Key words:  Distance Learning, E-education, Multi-Agent systems

The paper discuss a problem of a usage of virtual laboratories in teaching process at technical universities. The analysis of advantages and disadvantages of simulated laboratories and laboratories with an distant access is presented. Also different ways of implementing simulated virtual laboratories are given. Moreover, the problem of the supporting virtual laboratories by multimedia lessons and a usage of e-platforms is raised. Several examples of virtual laboratories application at Wroclaw Technical University are also given.
 

Key words:  E-learning, Virtual laboratories, Internet Technologies
 

FADI THABTAH
SModelling Optimisation Scheduling And Intelligent Computing Research Centre
Email: Fabdelja@bradford.ac.uk
 

PETER COWLING
Modelling Optimisation Scheduling And Intelligent Computing Research Centre
Email: P.i.Cowling@bradford.ac.uk

YONGHONG PENG
SDepartment of Computing, University of Bradford, BD7 1DP, UK
Email: Y.h.Peng@bradford.ac.uk
 

Association rule discovery is one of the primary tasks in data mining that extracts patterns to describe correlations between items in a transactional database. Using association rule mining for constructing classification systems is a promising approach. There are many associative classification approaches that have been proposed recently such as CBA, CMAR and MCAR. In this research paper, four associative rule algorithms (CBA, CMAR, CPAR, MCAR) have been compared with reference to accuracy against 12 benchmark classification problems. Our goal is to determine the most accurate technique in forecasting the future classes of unseen test data objects. After experimentation with different data sets, the results revealed that none of the investigated techniques dominated the others with regards to accuracy. Moreover, MCAR produced more accurate classification systems than CBA, CMAR and CPAR, respectively. This is due to the less pruning operation employed by MCAR, which leads to generating larger classifiers. A post pruning method is recommended to reduce the number of rules generated by MCAR, where it is obvious for cases like “Cleve” and “Germany” data sets.

Key words:  Association Rule, Classification, Data Mining, Associative Classification, Prediction Accuracy