Key Engineering Materials Vols. 460-461

Abstract: This paper mainly analyzed the principle of multi-source spatial data fusion, and expounded the multi-source spatial data fusion of the distributed model structure. The paper considers a distributed multi-sensor information fusion system factors, A performance evaluation model was established which was suitable for distributed multi-sensor information fusion system, It can estimate the system's precision, track quality, filtering quality, and the relevant between Navigation Paths and so on. Meanwhile, we had a lot of experiments by the datum which generated by the simulation test environment, experiments show that this evaluation model is valid.

Abstract: Apriori algorithm is one of the most classical algorithm in association rules, however, the algorithm is low efficiency, such as firstly it needs to repeatedly scan the database, which spends much in I/O. Secondly, it create a large number of 2- candidate itemsets during outputting frequent 2- itemsets. Thirdly, it doesn’t cancel the useless itemsets during outputting frequent k- itemsets. In the paper, it describes an improved algorithm based on the compressed matrices which improve the efficiency during creating frequent k- itemsets on three aspects, which simply scans the database once, after compressed transactional matrix, and by multiplied matrix get the frequent item sets, which effectively improved the efficiency in mining association rules.

Abstract: Pre-coding type can be grouped into two approaches, unitary or non-unitary for 3GPP Long Term Evolution(LTE). Output SINR of unitary Pre-coding for MU-MIMO is investigated in this paper. A more simple method of calculation for SINR is presented, which Leads to the computational complexity down to 18% compared with original method, while calculating MIMO Pre-coding system with 4 matrices (i.e. 8 vectors). In addition, we also analyzed the loss of SINR when using unitary or non-unitary Pre-coding. Corresponding simulations show that output SINR of non-unitary pre-coding system is inferior to the unitary pre-coding case and such losses can been described using statistical linear approximation.

Abstract: The development of the BP NN is part of the research project of internet based intelligent supplier selection and evaluation system. According to obtained information from Google Search Engine, the possible suppliers will be classified and put into the data file which is used for the further evaluation. In the paper the BP NN model used for supplied evaluation is introduced first, and then the six input variables have been discussed for the BP NN model. The raw data of each variable has to be converted into BP NN model accepted format for training and application. The conversion methods for each valuable have been discussed in the paper as well. Final, the test results show the successful system running.

Abstract: Business use of Internet has exposed security as one of the key-factors for successful online competition. There are very few sound existing ISG frameworks that can effectively guide most organizations in their information security governance endeavors. This paper presents a collaborative framework for information security management system in order to elucidate how information security should be addressed at an executive level. The advanced intelligent multi-agent technology is applied in this proposed framework. In order to evaluate the proposed framework, this paper describes a prototype system based on this proposed collaborative framework. This prototype system suggests that the proposed collaborative framework is feasible, correct and valid.

Abstract: With the growing population of using public network, network intrusion events happened more frequently. Aim to the drawbacks of traditional information security methods, the multi-agent system model was proved to be very promising. This paper proposes a dynamic distributed information security system based on multi-agent that dynamically integrates all distributed information security facilities to achieve efficiently collaborative information security. In order to validate this proposed approach, the authors give the material experiment test in the end of this paper. The final experiment result suggests that, this proposed approach is feasible, correct and valid.

Abstract: Collaborative filtering is very effective in recommendation systems. But the recently researches have proved the collaborative filtering is significant vulnerable in the face of profile injection attacks. Profile injection attacks can be identified to some attack models. The attacker can easily bias the prediction of the system based on collaborative filtering algorithms. In this paper, an improved algorithm based on Singular Value Decomposition is proposed. Some dimensions are chosen by the improved algorithm to find capture latent relationships between customers and products. In addition, the robustness of the algorithm is improved by the way. Several experiments are conducted. The results suggest that the proposed algorithm has advantages both in robust and stable over previous algorithms.

Abstract: The student early warning is a type of the problem that involves intelligence factor and non-intelligence factor and indetermination factor, the challenge lies in ascertaining early-warning factor. In this paper, basing on analyzing association rule and the principle of Apriori algorithm, a new Apriori algorithm has been designed for solving complexity of predictors in an early warning system of college students. At first, make sure the count K of item gathering the largest frequent itemset. Then directly make the largest frequent itemset LK by the items, whose count of item is larger than K or equal to K. We can acquire the k-1、k-2 frequent itemset with the same method. We carry on an experiment on the new algorithm. As a result, the efficiency of the new Algorithm is raised obviously. And along with data quantity aggrandizement, the time consumed changes inconspicuously. That is to say the algorithm keeps very high efficiency.

Abstract: Intrusion detection system is an important device of information security. This article describes intrusion detection technology concepts, classifications and universal intrusion detection model, and analysis of the intrusion detection systems weaknesses and limitations. Finally, some directions for future research are addressed.

Abstract: Wavelet analysis has become a popular subject in scientific research for twenty years. It has been a powerful tool for exploring and solving many complica -ted problems in natural science and engineering computation. The notion of exact orthogonal nonseparable four-variate wavelet packs is introduced. A novel method for constructing them is provided by iteration method. The orthogonality properties of four-variate wavelet packs are discussed. Three orthogonality formulas concerning these wavelet packs are provided. The pyramid decomposition scheme is derived