Applied Mechanics and Materials Vols. 602-605

Abstract: In the field of network security, logs analysis is an efficient protecting measure. But it couldn’t meet with the requirements of dealing with threaten rapidly and responding at real time only with security administrator’s analysis of abundant multidimensional IDS logs. Visualization of multidimensional data is applied in information cognition and decision, with people’s perception of graph and advanced technology of computer’s dealing with abundant data. A frame of visualization in intrusion detection is proposed according to the process of visualization and intrusion detection. And then a simple efficient attribute encoding is put forward. K-dimensional logs are converted to the right regular polygon of even sides (RPES) using optimization algorithm. The experiments show that this method offers more dimensional distribution information and implicit knowledge and help security administrator to make efficient decision.

Abstract: In this paper, we combine the complex network theory and the traditional software structure metrology to propose a new model for the study of the structural characteristics of the software---- Multi-dimensional measurement model of the software structure properties. The multi-dimensional measurement model of the software structure properties is divided into three parts. Each part has their own properties. In this model, the system is abstracted into a network model in the first step. Then we design the metric parameters considering both the complex network theory and the Object-Oriented software research and also give the definition and calculation method of these metric parameters. And on this basis we use the advantages of eclipse and complex network simulation tool pajek to calculate the metrics parameters designed before. Then give explanation to the experimental results which can demonstrate the reliability of the new model, which has also made a solid foundation for the following study of the software structure properties.

Abstract: The popular approaches for face recognition are PCA and LDA methods. But PCA could not capture the simplest invariance unless information is explicitly provided in the training data and LDA approach suffers from a small size problem. 2DPCA could reduce high dimensional data to a low-dimensional space.2DLDA could extract the proper features from image matrices based on LDA. Ensemble incomplete wavelet analysis method for face recognition is proposed based on improved fuzzy C-Means in this paper. The method proposed shows that it improves the accuracy and reduces the running time.

Abstract: With the characteristics of timeliness, rapid spread and easy access, micro-blog has made it possible to timely unearth emergencies and dynamically track the latest development. Considering frequency and timeliness of the words as well as the influence of resources, this paper proposes a sudden topic detection method, especially for micro-blog, and establishes a context-based weight evaluation model. Compared with semantic similarity model, this context-based model is especially more adaptable to micro-blog. Real experiment data from micro-blog proves that the proposed methods could detect sudden incidents effectively with big data processing capacity and low time complexity.

Abstract: The design and implementation of the Remote Monitor System for Simulation Elevator Based on EPA were appeared. The Remote Monitor System consists of the device layer, control layer, the server layer and client layer, the monitoring software, configuration software and so on, which implements to the simulated elevator data acquisition and intelligent control of output and remote monitoring.

Abstract: An acquisition system applying in transient electromagnetic method (TEM) receiver has been designed in this paper. The accuracy of acquisition has been improved by using a floating-point A/D with high input dynamic range, which is composed of programmable gain amplifiers (PGA) and a dual-channel 16 bit resolution ADC. The real-time performance can be ensured by using a field-programmable gate arrays (FPGA) as the main control unit to control ADC, PGA, data storage and communication interface. By integrating a soft core CPU, Nios II, in FPGA, the complexity of circuit design was decreased, and the work efficiency of system was improved. This system, which has good performance of real-time and high integration level, can meet the needs of TEM signal acquisition.

Abstract: Due to the large amount of network data and complex representation, traditional network security behavior recognition system always leads to high redundancy and dimension, resulting in taking up more resources, larger computation. To solve this problem, we do the features selection. This article presents a consensus decision-making method, which combines current famous feature selection algorithms to obtain a more reasonable result and to sort the features in order of importance to facilitate the appropriate selection of features under different conditions. With this method tested on SVM (Support Vector Machine) as classification algorithm, it proves that the algorithm effectively improves the recognition accuracy with fewer features and performs better in terms of result stability.

Abstract: In order to realize the mine water source identification and to prevent water accidents, taking Xinglong Zhuang coal mine’s lower group coal mining as an example, based on the analysis of mine hydrogeological conditions, the mine water source identification system was designed by using the software of MapObjects powerful spatial analysis ability and Bayesian classification analytical method. The software can achieve mine water source identification and provide a certain basis for mine water disaster prevention and control.

Abstract: A stayed cable inspection system was developed which consists of robot, host computer, cameras and image acquisition system. The robot was driven with single motor and could climb cables of various and variable diameters. Pictures of the cables’ were taken by the robot, and the defects and mars were identified automatically with image recognition. The steps of image recognition includes image de-noising, image enhancement, image segmentation, feature extraction, and recognition with the features of the images’ histogram grayscale distributions and energy distributions.

Abstract: HSL (HSL line) noise mainly exhibits on the noise source and its radiation intensity. Testing on noise can be done at the location with maximal noise vertical distribution, by monitoring on certain points continuously 14h, general monitoring to equivalent sound lever of 1h operation no less than average density day and night, and independent monitoring to noise susceptible architecture.