Applied Mechanics and Materials Vols. 543-547

Abstract: This paper purposes a K-means clustering algorithm based on improved filtering process. Thealgorithm improves the filtering process,The two minimum sample points are reasonable initial clustering centers. It makes the probability summary of data in a cluster as large as possible, and the probability summary of data in different clusters as small as possible. Experimental results show that the proposed algorithm can select the proper initial clustering center, and it is more compact and robust than thetraditional K-means clustering algorithm.

Abstract: The automatic dependent surveillance-broadcast (ADS-B) system is the backbone of the next-gen air traffic control (ATC) modernization plan. Unfortunately, ADS-B system suffers from serious cyber-security vulnerabilities due to the open broadcast of aircraft data, without regard to message confidentiality. However, using common encryption scheme to provide confidentiality of ADS-B data is not a good solution, because encrypting data with ordinary cryptosystem would violate the original openness intention of ADS-B system design. In this paper, based on the new format-preserving encryption (FPE), we present an efficient data encryption scheme for the ADS-B data. The security analysis demonstrates that our scheme can achieve confidentiality of ADS-B data. The performance evaluation shows that the scheme is computationally efficient for the typical avionics devices with limited resources.

Abstract: With the continuous expansion of computer simulation scale, the demand for data mining algorithm is also more and more big. The difficulties in computer data mining technology are focused on algorithm development. Apriori algorithm is a kind of computer data mining algorithm which can greatly improve the computational efficiency. The algorithm uses association rule, which can avoid repeated frequently by layer scanning, reducing the computer time. This paper uses Apriori algorithm to design the data mining parameter optimization model of computer 3D human biology simulation, and applies to improve the step three jump. Through the simulation we found step distance appropriate, it provides technical reference for the application of computer simulation technology in sports.

Abstract: With the rapid development of network and database technology, data need to be processed massively increased, how to carry out effective data mining is a serious problem. The mature development of granular computing algorithm provides new ideas and new methods to study for data mining. Association rules of granular computing can reduce the number of object scanning data set, and improve the efficiency of the algorithm. In this paper we introduce the data source, classification, technology, system structure, operation process, application in other areas of data mining technology. Based on association rules of granular computing, data mining technology can provide quantitative basis for enterprise in screening assessment, so the service object has a stronger competitive advantage and focus more on its problems.

Abstract: In case of experimental data contaminated with errors and noise, the robust ε-support vector regression has good forecast accuracy and high generalization ability. However, it depends on the selection of system parameter. Firstly, this paper introduces the robust ε-support vector regression method. Secondly, as the experiments prove, the new method achieves high forecast accuracy by virtue of the optimal penalty parameter C. Finally, the optimal method of parameter C is presented in the last section.

Abstract: The key to the robust ε-support vector regression algorithm is searching for the optimal regression hyper plane while data with disturbance in the X-direction. In the paper, the optimal regression hyper plane and the optimal separating hyper plane are compared and analyzed. By means of Kolmogorov test, it is can be deduced that the testing errors of the robust ε-support vector regression experiments follow normal distribution. The result demonstrates that the algorithm has good forecast accuracy and high robustness.

Abstract: With development of Web2.0, user-generated-contents spread in the Internet. It provides good topics to the research. People express their opinions and sentiments on the cyberspace. The opinions and sentiments are very important and attract extensive research. However, it is impossible for the user to browse the content carefully. Hence classification and summarization of online text become a pressing issue. In this paper, we propose a Co-Training method to the sentiment classification. Posts and replies have been chosen as different views for the Co-Training. Several features have been employed. Experimental result in the datasets demonstrates the advantage of the proposed model.

Abstract: Based on LabVIEW, the comprehensive evaluation system of armored vehicle cabin environment is designed and developed. Combining the function and demand of the evaluation system, the system evaluation process is proposed and the system general structure is designed. Through the method of dynamic calling subvi on subpanel, hybrid programming with MATLAB and LabVIEW, and the organic combination of the For Loop Structure and form controls, the specification of software interface design, data input/output and weighted filtering analysis of test data in the time-frequency domain and other technical problems are solved.

Abstract: Token bucket algorithm is a common algorithm used for traffic shaping and rate limiting. In LTE (Long Term Evolution) system, RRC (Radio Resource Control) layer configures priority, PBR (prioritisedBitRate), and BSD (bucketSizeDuration) for every logical channel so that lower priority channel can also get resource and bursting data can be sent in a short time. This paper firstly introduces application of token bucket algorithm in LTE system. And the test procedure is designed to test the algorithm. Test result shows that the algorithm is used in LTE system successfully. The algorithm has been used in TD-LTE radio frequency conformance testing instrument.

Abstract: A bootloader's main function is to initialize the hardware, pass an abstraction of the initialized hardware, a hardware description, to and execute the kernel. While most bootloaders concentrate on ARM or PowerPC architecture, in this paper we propose a lightweight bootloader based on MIPS architecture. The lightweight bootloader is mainly for mobile devices. First, we introduce the architecture of the bootloader and some special requirements for mobile devices are considered. Then, based on the architecture, we implement the bootloader in MIPS assembly language and C language. To validate our research, we test the bootloader in a real MIPS-based evaluation board. The results show that our bootloader works quite well. With MIPS technology's role becoming more and more important in mobile Internet, we hope that the bootloader developed in this paper will be utilized in the future.