2008 Second International Symposium on Intelligent Information Technology Application

IP traffic prediction plays an important role in network-layout, traffic-management, as well as the emphasis of traffic-project, congestion-control and network management. Poor prediction performance would be acquired generally as a result of intense nonlinearity of networks traffic. To tackle it, a modeling method for exact representing IP trafficpsilas movement tendency and a regression algorith... View full abstract»

A popular and effective method for interpolation of data is by means of radial basis functions, but the most significant drawback of this method, which made researchers unwilling to use it for practical purposes, is its severe computational cost. Introduction of different methods by the researchers active in this field, however, has alleviated this problem. This paper presents a new method for red... View full abstract»

According to the current research on the bus passengers flow counting, this paper proposes a new counting method of bus passenger flow based on kinematics of human body and SVM. This method solves the problem that low cost device canpsilat count the number of passengers accurately and canpsilat distinguish the direction of passengerspsila movement. Finally, this paper proves the feasibility and ef... View full abstract»

Existing algorithms that mine graph datasets to discover patterns corresponding to frequently occurring sub-graphs can operate efficiently on graphs that are sparse, contain a large number of relatively small connected components, have vertices with low and bounded degrees, and contain well-labeled vertices and edges. However, for graphs those do not share these characteristics, these algorithms b... View full abstract»

Collaborative filtering recommendation system is a widely used method of providing recommendations using explicit ratings on items from users, which provides personalized recommendations on products or services to customers. However, the current research on recommendation has paid little attention to the use of time-related data in the recommendation process and the study on collaborative filterin... View full abstract»

Recommender systems are becoming increasingly popular with the evolution of the Internet, and collaborative filtering that using explicit ratings on items from users is the most successful technology for building recommendation systems. But traditional collaborative filtering algorithm is not suitable for itempsilas multiple content and multiple level recommendations. So, a new concept hierarchy m... View full abstract»

Analyses were made on the basic principles of evolutionary algorithm, evolution strategies and evolution programming. Considering the superiority of evolutionary algorithm in intellectual computing, we analyze a typical optimizing algorithm for artificial immune system (AIS). Combining evolutionary algorithm and artificial immunity, we present an immune intrusion analysis scheme based on statistic... View full abstract»

This paper proposes an invertible semi-fragile video watermarking algorithm using a hash function to authenticate the MPEG-4 video contents. The proposed algorithm embeds two watermarks into I frame while encoding YUV video to MPEG-4 format. One watermark using hash function aims to authenticate the contents and embed the frame number for manipulation location between frames, and the other one bas... View full abstract»

A novel method for forecasting the silicon content in hot metal is proposed using principal component analysis (PCA) and BP network. PCA can consider the correlations among multiple quality characteristics to obtain uncorrelated principal components. These principal components are then taken as the input parameters of the BP neural network. Then the BP network models are established and trained to... View full abstract»

One of the main difficulties in machine learning is how to solve large-scale problems effectively, and the labeled data are limited and fairly expensive to obtain. In this paper a new semi-supervised SVM algorithm is proposed. It applies tri-training to improve SVM. The semi-supervised SVM makes use of the large number of unlabeled data to modify the classifiers iteratively. Although tri-training ... View full abstract»

A method is presented to obtain maps between the high-dimensional data and the low-dimensional space deduced by locally linear embedding (LLE). Since LLE does not provide a parametric function that build maps between the image space and the low-dimensional manifold. In this paper, multivariate linear regression is applied to deduce the maps. It can successfully project a new data point onto the em... View full abstract»

This paper presents an improved image retrieval algorithm which based on the gray projection cyclostyle matching to overcome the shortcomings of conventional content-based image retrieval methods. The algorithm based on the analysis that gray scale projection curves of horizontal and vertical direction of the images fully reflects the feature of gray distribution in the corresponding direction, th... View full abstract»

Diverse pose estimation of three-dimensional (3D) object in the whole view-space remains a challenge in the field of pattern recognition. In this paper, a pose estimation algorithm of 3D object named isomap-eigenanalysis-regression (Isomap-E-R), which estimates arbitrary pose of 3D object in the whole view space, is proposed. For the training set, the low-dimensional embedding of input pattern set... View full abstract»

Recommendation systems need to employ efficient reality mining module to find out whether social relationships extracted from a collected dataset are really meaningful in real world. Also, the recommendation should be generated by taking current contexts (e.g., location and time) account into. In this paper, we want to propose an ontology-based framework to discover real social networks from a lar... View full abstract»

In the text analysis, the Dirichlet compound multinomial (DCM)distribution has recently been shown to be a good model for documents because it captures the phenomenon of word burstiness, unlike the standard multinomial distribution. In this paper, for the sake of improving performance of modeling documents, we propose a variant of DCM and Gibbs distribution called Dirichlet multinomial Gibbs (DMG)... View full abstract»

This paper studies hybrid dynamical evolutionary algorithm in the context of classification rule discovery. Nature inspired search algorithms such as genetic algorithms, Ant colonies and particle swarm optimization have been previously studied on data mining tasks, in particular, classification rule discovery. We extended this work by applying a hybrid algorithm which combines dynamical evolutiona... View full abstract»

To obtain a high robust of speech recognition for noisy conditions, a new pre-processing stage based on wavelet thresholding algorithm is proposed in this paper. The purpose of using the DWT is to benefit from its localization property in the time and frequency domains. Compromise function is proposed compared with hard and soft thresholding function. A new thresholding value, Neyman-Pearson crite... View full abstract»