Applied Mechanics and Materials Vols. 268-270

Abstract: In order to reduce times of software regression testing, a new research idea and method is proposed based on RBFN (Radial Basis Function Network). Using the adaptive ability of network study, regression testing is optimized by its learning strategy. The simulation results demonstrate the new method can forecast regressive testing effectively, and implement very little error. It means an important meaning for developing new effective method of soft testing in the future.

Abstract: Based on analyzing EPON Network architectures and EPON Network Management functional requirements, this paper designed an EPON Network management System according to configuration management, performance management, fault management, topology management, security management and log management, and achieved a system based on RIA technology and B/S mode. The client was written based on Flex framework and the server was written by JAVA. A middleware technology was used between client and server for communicating. Through the B/S mode and SNMP, the system has achieved a series of management functions of EPON network.

Abstract: Image compression is a common operation in the digital image processing. In order to get high compression rate while prevent blurred edges, an image compression algorithm based on multilevel B-Spline is proposed. Firstly, image plane is seen as the parameter domain of B-Spline and gray values are seen as value of B-Spline. Then, gray values are approximated by B-Spline iterately and at each iteration control points are double dense of the former iteration. The error criterion used is an absolute one and different tolerances are used for edge points and common points, which make edges still clear in the compressed image. Experiments demonstrate that the algorithm presented guarantees that gray value loss at each pixel is less than a preset tolerance while the edge feature of original image is preserved perfectly.

Abstract: Decision tree algorithms have been widely used in intrusion detection. In this paper, within the framework of granular computing (GrC), we propose a new decision tree algorithm called DTGAE and apply it to intrusion detection. First, by virtue of the GrC model using information tables, we propose a new information entropy model, which contains two basic notions: approximation entropy of granule (AEG) and GrC-based approximation entropy (GAE), where the latter is defined based on the former. In algorithm DTGAE, GAE is adopted as the heuristic information for the selection of splitting attributes. When calculating AEG and GAE, we not only utilize the concept of conditional entropy in Shannon's information theory, but also use the concept of approximation accuracy in rough sets. Second, we present a method of decision tree pre-pruning based on Düntsch's knowledge entropy. Finally, the KDDCUP99 data set is used to verify the effectiveness of our algorithm in intrusion detection.

Abstract: Although the value of factor weight in an evaluation work is deterministic, the solving process is random, so connection between weight solution with digital characteristics or distribution functions of specific random variables or random process could be build. Using stochastic simulation method to get a lot of random solutions to the problem, expectation of the random solutions can be used as a estimation solution. On basis of idea of Monte Carlo simulation, this paper analyzed the probability process of calculating factor weight, and provided the procedures of estimating factor weight by means of Monte Carlo simulation. Through discussion and example in this paper, feasibility and validity of this method were proved, which may make foreshadowing for follow-up research work.

Abstract: Using wireless communication technology, location-based advertising (LBA) is booming recently. Most LBA researches are cross-sectional analysis studying only contemporaneous impacts, which might seriously underestimate advertisings’ efficacy in the long run. Complementary to these researches, we employ a company archival dataset to explore LBA’s and PUA’s (Pop-up Advertising) sales impact in both contemporaneous and long run terms. The developed dynamic structural equation model reveals that (1)in terms of magnitude, LBA’s efficacy is stronger than that of PUA both contemporaneously and accumulatively; (2) in terms of persistent time, while LBA only have contemporaneous impact, PUA’ s impact lasts for 9 days. Theses empirical findings would help managers to allocate advertising budget efficiently while using wireless communication technology.

Abstract: Along with the development of multimedia technology, large capacity image library a wide range of applications, content-based image retrieval (CBIR) technology is management and retrieval effective means. Among the contend-based retrieval techno-logiest, feature extraction is most important. For instance, color, texture and shape feature etc. But each feature of image can only catch one aspect of the similarity of image, how to represent images better has become a important research field in content -based image retrieval. Introduced the use of image texture characteristics of image retrieval method, with gray symbiotic matrix algorithm, from the block of the texture characteristics, and gives the CBIR system is realized, and the experiment results are given.

Abstract: With the development of database technology as well as the widespread application of database management system, the capability of collecting data was improved rapidly, and lots of data have been accumulated. Data mining was created for the purpose of excavate the useful knowledge hidden behind these data. Data classification is not only an important issue of data mining but also an effective KDD method. Decision Tree, which is a major technology of data classification, is applied far and widely. In this article, the concrete step of mining data by decision tree, the main algorithm and the basic idea of decision tree were summarized and analysed.

Abstract: Customer demand is the driving force behind the development of enterprises. It is the core of mass customization production mode that access to customer demand information, analysis and management. Questionnaire is an important way to get customer demand. This paper presents a design of closed-end questionnaire about electric bicycle which has two characteristics: (1) highlighted the access to subject features of customer. (2)The numerical answers will be processed using fuzzy logic.

Abstract: This paper is intended to explore the weakness of Li et al.’s billing service user authentication scheme. The users of this authentication scheme are required to compute a series of session keys and use different session key in different session to strengthen the robustness of this scheme. They are not only difficult to remember but also can be figured out by attacks, making authentication function useless. In addition, the system may cause denial of service because synchronization could not always be guaranteed. We illustrate all possible attacks against this system that may be realized, and give an analysis of its weakness.