In order to improve the performance of Chinese text categorization, a new Chinese text categorization method based on angle distribution is presented. The new method describes the text with a more precise model and pr...
详细信息
In order to improve the performance of Chinese text categorization, a new Chinese text categorization method based on angle distribution is presented. The new method describes the text with a more precise model and proposed a new categorization algorithm by employing angle distribution. Simulation results on open Chinese text collection show that the precision and recall of most classes have been increased with reference to the classical method, and the macro average of precision and recall are both about 72 percents, which certificating the effectiveness and feasibility of the angle distribution-based algorithm.
Computational grids aim to aggregate the power of heterogeneous, geographically distributed, multiple-domain spanning computational resources to provide high performance or high-throughput computing. To achieve the pr...
详细信息
Computational grids aim to aggregate the power of heterogeneous, geographically distributed, multiple-domain spanning computational resources to provide high performance or high-throughput computing. To achieve the promising potentials of computational grids, an effective and efficient scheduling system is fundamentally important. Grid is also a dynamic resource-sharing computing environment in which there are many applications running. At time of scheduling decision making, accurate resource predication on run-time resource measurement parameters is critical to maximize application performance. In this paper, a method for estimating task running time and two scheduling algorithms are presented to facilitate effectively dynamic scheduling in a heterogeneous grid environment and improve application performance.
In this paper, the problem of ranking sentence pairs in parallel corpora was addressed for the first time. To solve this problem, a novel model was proposed. In this model, both syntax features and semantics features ...
详细信息
In this paper, the problem of ranking sentence pairs in parallel corpora was addressed for the first time. To solve this problem, a novel model was proposed. In this model, both syntax features and semantics features of sentence pairs are considered. Since most today's statistical machine translation models depend on word alignment, features related to word alignment information are also included. Two experiments were carried out and the results showed that the model had promising performance.
In this paper, we present the necessary and sufficient conditions of two finite classes of samples that can be separated by a hyperplane in terms of support vectors which are just the vertices of a convex hull of each...
详细信息
In this paper, we present the necessary and sufficient conditions of two finite classes of samples that can be separated by a hyperplane in terms of support vectors which are just the vertices of a convex hull of each class of samples. We also extend the calculating formula of the margin of an optimal separating hyperplane to some cases of the classes of infinite samples in Hilbert space. These results are the generalization and improvement of the corresponding results for the theory of SVM in Euclidian space.
A small scale distributed computing system that is able to meet the needs of parallel intelligent techniques for engineering and science applications is reported in this paper. The reported system is a cluster of gene...
详细信息
A small scale distributed computing system that is able to meet the needs of parallel intelligent techniques for engineering and science applications is reported in this paper. The reported system is a cluster of general-purpose PCs interconnected in a network. Such systems are powerful yet low cost. While such systems are not new, most applications written for cluster systems are programmed in MPI-C. Results from the proposed system have shown that it is advantageous to use Java due to its simplicity and mobility. Hence, the goal of this research is to design a Java-based parallel platform for efficient execution of intelligent techniques for engineering and science applications on a cluster system.
The efficiency of transferring a pack of small-sized flies or a large-sized file is very low in FTP based file transfer system. To solve this problem, we have studied the optimized method of different types of file tr...
详细信息
The efficiency of transferring a pack of small-sized flies or a large-sized file is very low in FTP based file transfer system. To solve this problem, we have studied the optimized method of different types of file transfer and present an adaptive transfer algorithm. Based on this algorithm. We design and implement the file transfer subsystem of global distributed storage system.
Peer-to-peer network can be seen as truly distributed computing system. Each peer is both a client and a server in these networks. The trust between peers cannot be set up simply on the traditional trust mechanism. A ...
详细信息
Peer-to-peer network can be seen as truly distributed computing system. Each peer is both a client and a server in these networks. The trust between peers cannot be set up simply on the traditional trust mechanism. A reasonable trust construction approach comes from the social network analysis. Trust relations between individuals are set up upon recommendations of other individuals. In this paper, we propose a recommendation-based global trust model for peer-to-peer network, which is easy to implement. Then we introduce some security mechanism into this model to defense several attacks, such as tamper, pretend, slander and exaggerate.
In the fuzzy reasoning of expert system, similarity-based fuzzy reasoning method is a kind of simple and important method. This paper indicates the limitation of an existing similarity measure and its algorithm, and p...
详细信息
In the fuzzy reasoning of expert system, similarity-based fuzzy reasoning method is a kind of simple and important method. This paper indicates the limitation of an existing similarity measure and its algorithm, and proposes an improved similarity measure and reasoning method. We give an example and comparison of two methods to show the superiority of the new algorithm.
This paper investigates convergence theorems of (HNNMDs). (The discrete Hopfield network with multiple delays). We have demonstrated how to transform a network with multiple delays to a new network with a single delay...
详细信息
This paper investigates convergence theorems of (HNNMDs). (The discrete Hopfield network with multiple delays). We have demonstrated how to transform a network with multiple delays to a new network with a single delay. Based on serial rule updating mode, the conditions for convergence of a HNNMDs can be obtained.
A good intrusion detection system (IDS) should have high precision on detecting attacks and low false alarm rates. machinelearning techniques for IDS usually yield high false alarm rate. In this work, we propose to c...
详细信息
A good intrusion detection system (IDS) should have high precision on detecting attacks and low false alarm rates. machinelearning techniques for IDS usually yield high false alarm rate. In this work, we propose to construct host-based IDS for flooding-based denial of service (DoS) attacks by minimizing the generalization error bound of the IDS to reduce its false alarm rate and increase its precision. Experiments using artificial datasets support our claims.
暂无评论