Scientia Iranica
Volume:17 Issue: 1, 2010

Data mining is a widely used approach for the transformation of large amounts of data to useful patterns and knowledge. Fuzzy association rules mining is a data mining technique which tries to nd association rules without the e ect of sharp boundary problems when data contains continuous and categorical attributes. Grid data mining is a new concept, which allows the data mining process to be deployed and used in a data grid environment where data and service resources are geographically distributed. In this paper, a grid service for mining fuzzy association rules is developed. The service is implemented based on recently proposed Data Mining Grid Architecture (DMGA) and uses the Web Service Resource Framework (WSRF). Experimental evaluations, after implementing and deploying the service, show the e ectiveness and acceptable performance of the proposed grid service. Additionally, in this study, a new algorithm, namely FFDM, is developed to mine fuzzy association rules without raw data exchange, using the distributed storage of data grid environments. Empirical evaluation of FFDM reveals the scalability and eciency of the proposed method, in addition to the advantages of minimum messaging and providing privacy of data.

The signal integrity problem, especially crosstalk noise, is an important issue in physical design in the nanometer regime. Wire multiplexing is a recently proposed method to reduce these problems in current design methodologies. This technique is presented to increase the routability of the design and also reduce crosstalk noise by serializing parallel wires via delay insensitive asynchronous serial transceivers. In this paper, this technique is improved in terms of routability and computation time and, also, its impact on crosstalk is examined. Finally, it is evaluated with 180 nm and 130 nm technologies. Experimental results show that for the attempted benchmarks, congestion and routability are improved by 15.54% and 21.57% on average, respectively, and crosstalk noise is improved by 9.51% on average. These improvements are achieved at the cost of a slight increase in power consumption (0.12% on average) and runtime (less than 10.01% on average).

The Session Initiation Protocol (SIP), which is becoming the de facto standard for the next-generation VoIP networks, is currently receiving much attention in many aspects. One aspect that was not deeply addressed in the original SIP is its authentication procedure. Apart from its security, an SIP authentication procedure should be ecient. This paper proposes a robust and ecient three-party SIP authentication protocol. In this protocol, the end users are authenticated with the proxy server in their domain using the registrar server. Compared to previous works, our proposed protocol is more ecient and secure. To support our protocol with a formal security proof, its model is constructed using High-Level Protocol Speci cation Language (HLPSL). The model is veri ed using the model checking tool, AVISPA, and the result con rms that the protocol is quite safe.

In this paper, we study a new simple method to nd a numerical solution for the System of Non-linear Di erential Equations (SNDE). The method is based on a matrix form of SNDE and we use computational aspects of special matrices to nd the expansion of the unknown function. Finally, some numerical examples illustrate the accuracy of the method.

This paper implements models of two important distributed generators. The rst developed model describes and simulates the dynamic performance of a Single Shaft Micro Turbine, while the second model simulates the dynamic behavior of a Solid Oxide Fuel cell (SOFC). Increasing and reducing the power produced by these micro sources is also simulated. The two developed models can test the dynamic response for the power demand change from 0% to 120% of the rated value. The stand alone dynamic performance of the two developed models is analyzed and evaluated. Results proved the e ectiveness and robustness of the developed models. The two developed models can be used to describe the behavior of a Micro Grid (MG) under di erent disturbance conditions like islanding from the main grid, load following, load shedding, unbalanced loads, failure of one micro source and so forth. Also, we developed models for those micro sources for building a complete model that can describe the overall dynamic performance of the Micro Grid (MG). The developed models of the Micro Turbine and the SOFC are inserted in a complete MG system to prove the validation of those models. Investigations of the MG system under di erent disturbance conditions are performed. Results proved the validation and e ectiveness of the two developed models in the studying and analysis of the transient dynamic response of MG. All models and controllers are built in the Matlab r Simulink r environment.

In this article, the interaction of an arbitrary number of quantum dots behaving as arti cial molecules with di erent energy levels and a multi-mode electromagnetic eld is studied. We make the assumption that each quantum dot can be represented as an atom with zero kinetic energy, and that all excitonic e ects except dipole-dipole interactions may be disregarded. We use the Jaynes-Cummings-Paul model with applications to quantum systems based on a time-dependent Hamiltonian and entangled states. We obtain a system of equations describing the interaction, and present a method to solve the equations analytically for a single mode eld within the Rotating-Wave Approximation. As an example of the applicability of this approach, we solve the system of two, two-level quantum dots in a lossless cavity with two modes of electromagnetic eld. We, furthermore, study the evolution of entanglement by de ning and computing the concurrency.

Determination of hypnotizability is important, before prescribing any hypnotic treatment. Existing methods for measuring the level of hypnotic susceptibility are subjective, with some problems. In this study, a feature based on EEG weighted regional frequency was introduced, which can characterize the level of the subject''s hypnotizability objectively. The ability of this feature for making a signi cant di erence between three hypnotizable groups at the end of hypnotic suggestion was shown using statistical analyses. This feature was calculated based on the empirical mode decomposition method and the Hilbert transform. The EEG signals that were used in this study were recorded during hypnotic suggestion from 32 subjects. A K-nearest neighborhood-based classi er was designed for classi cation of the hypnotizable groups. The performance of the classi er was validated using the leave-one-out method, which showed the mean error of 3.13% in determination of the subject''s hypnotic susceptibility level. This evaluation and obtaining the error were done by comparing the new method''s results with the score of hypnotizability that was determined for each subject, using the subjective Waterloo-Stanford criterion. The new method, as opposed to common subjective clinical methods, represents a real time and objective procedure for determining hypnotic susceptibility.

This paper presents a family of Variable Step-Size (VSS) Ane Projection (AP) adaptive ltering algorithms with Selective Partial Updates (SPU) and Selective Regressors (SR). The presented algorithms have good convergence speed, low steady state Mean Square Error (MSE), and low computational complexity features. The stability bounds of the family of SPU-APA, SR-APA and SPU-SR-APA are analyzed, based on the energy conservation arguments. This analysis does not need to assume a Gaussian or white distribution for the regressors. We demonstrate the good performance of the proposed algorithms through simulations in system identi cation and acoustic echo cancellation scenarios.