检索结果-内蒙古大学图书馆

2011 International Conference on Asian Language Processing, IALP 2011

作者： Liu, Meiling Ren, Honge Zheng, Dequan Zhao, Tiejun Department of Computer Science and Application Northeast Forestry University Harbin China Machine Intelligence and Translation Laboratory Harbin Institute of Technology Harbin China

ISBN: (纸本)9780769545547

This paper introduces a model to describe the dynamic evolution of network information, identifying and analyzing the document collection on the same topic in different stages. In order to characterize the dynamic relationship of evolutionary content differences, this paper presents a dynamic multi-document summarization model, which is called the Dynamic Manifold-Ranking Model (DMRM). Some experiments were conducted on the Update Task test data from TAC2008, and results of new model were compared with results from the TAC2008 evaluation. This comparison demonstrated the effectiveness of the model. © 2011 IEEE.

关键词： Dynamic Manifold-Ranking Summarization dynamic dynamical evolution network information

来源：评论

学校读者我要写书评

暂无评论

Learning from imbalanced data sets with a Min-Max modular support vector machine

引用

中国电气与电子工程前沿 2011年第1期6卷 56-71页

作者： Bao-Liang LU Xiao-Lin WANG Yang YANG Hai ZHAO Center for Brain-Like Computing and Machine Intelligence Department of Computer Science and EngineeringShanghai Jiao Tong UniversityShanghai 200240China MOE-Microsoft Key Laboratory for Intelligent Computing and Intelligent Systems Shanghai Jiao Tong UniversityShanghai 200240China Center for Brain-Like Computing and Machine Intelligence Department of Computer Science and EngineeringShanghai Jiao Tong UniversityShanghai 200240China Department of Computer Science and Engineering Shanghai Maritime UniversityShanghai 201306China

Imbalanced data sets have significantly unequal distributions between *** between-class imbalance causes conventional classification methods to favor majority classes,resulting in very low or even nO detection of minority classes.A Min-Max modular support vector machine(M3-SVM)approaches this problem by decomposing the training input sets of the majority classes into subsets of similar size and pairing them into balanced two-class classification *** approach has the merits of using general classifiers,incorporating prior knowledge into task decomposition and parallel *** on two real-world pattern classification problems,international patent classification and protein subcellar localization,demonstrate the effectiveness of the proposed approach.

关键词： imbalanced data Min-Max modular network(M3-network) prior knowledge parallel learning support vector machine (SVM)

来源：评论

学校读者我要写书评

暂无评论

Fuzzy Classifiers - Opportunities and Challenges

引用

5th International Conference on Scalable Uncertainty Management (SUM 2011)

作者： Ralescu, Anca Visa, Sofia Machine Learning and Computational Intelligence Laboratory School of Computing Sciences and Informatics University of Cincinnati Cincinnati OH 45221-0030 United States Department of Computer Science College of Wooster Wooster OH 44691 United States

ISBN: (纸本)9783642239625;9783642239632

Several issues arise when we consider building classifiers in general, and fuzzy classifiers in particular. These issues include but are not limited to attribute/feature selection, adoption of a specific approach/algorithm, evaluate the classifier performance, etc. We consider the opportunities that such classifiers have to offer and contrast them with the challenges they pose.

关键词： classifiers fuzzy sets attribute selection error models

来源：评论

学校读者我要写书评

暂无评论

Retinal vessel tortuosity evaluation via Circular Hough Transform

Retinal vessel tortuosity evaluation via Circular Hough Tran...

引用

Iranian Conference of Biomedical Engineering (ICBME)

作者： Farnoosh Ghadiri HamidReza Pourreza Touka Banaee Morteza Delgir Computer Engineering Department Machine Vision Laboratory Ferdowsi University of Mashhad Iran Computer Engineering Department Machine Vision Laboratory University of Ferdowsi Mashhad Iran Ophthalmic Research Center Khatam-Al-Anbia Hospital Medical science University of Mashhad Iran Computer Science Department Robotics Laboratory Northeastern University USA

Retinal vessel tortuosity has shown to be significantly associated with cardiovascular diseases such as hypertension and diabetes. Despite importance of this field a few techniques have been proposed yet. All previous methods rely on a vessel extraction phase, which its accuracy affects final output and also time consuming. Nobility of presented algorithm is to introduce a method for evaluating retinal vessel tortuosity without any explicit vessel detection. We use the Circular Hough Transform (CHT) based on gradient field of the retinal image. Each vessel curve is detected as a semi-circle by Hough transform and tortuosity of the curve is determined with the help of accumulated value of circle center and its radius. As there are no any specific database for tortuosity evaluation, the algorithm was tasted on database consisting of 40 images, mixture of DRIVE database and images from Khatam-Al-Anbia Hospital consisting of 40 retinal images, of which 20 were tortuous and 20 were non-tortuous. The proposed algorithm can achieve classification rate of 92% along with less computation time in compare of previous methods.

关键词： Transforms Retinal vessels Classification algorithms Biomedical measurements Biomedical imaging Image segmentation

来源：评论

学校读者我要写书评

暂无评论

Robust low-rank subspace recovery and face image denoising for face recognition

Robust low-rank subspace recovery and face image denoising f...

引用

IEEE International Conference on Image Processing

作者： Mingyang Jiang Jufu Feng Key Laboratory of Machine Perception MOE Department of Machine Intelligence School of Electronics Engineering and Computer Science Peking University Beijing China

We propose a low-rank subspace recovery and image denoising method for face recognition. Traditional subspace methods commonly assume that face images from a single class lie on a low-rank subspace. However, due to shadows, specularities, occlusion and corruption, real face images seldom reveal such low-rank structure. To address this problem, we cast the problem of recovering face subspace from noisy images as a problem of recovering a low-rank matrix with sparse error of arbitrary large magnitude. By using the recent breakthroughs in convex optimization, we can exactly recover the subspaces from corrupted facial data. We apply this method to two well-known subspace methods: nearest subspace and sparse representation face recognition. The results show that our method is efficient in recovering the low-rank face subspaces by removing the noise in the training images, thus significantly improve the robustness of these methods in the presence of occlusion and corruption in both train and test images.

关键词： Face Training Noise Face recognition Robustness Sparse matrices Image denoising

来源：评论

学校读者我要写书评

暂无评论

Research on Multi-document Summarization Model Based on Dynamic Manifold-Ranking

Research on Multi-document Summarization Model Based on Dyna...

引用

International Conference on Asian Language Processing (IALP)

作者： Meiling Liu Honge Ren Dequan Zheng Tiejun Zhao Machine Intelligence & Translation Laboratory Harbin Institute of Technology Harbin China Department of Computer Science and application Northeast Forestry University China Department of Computer Science and application Northeast Forestry University Harbin China

关键词： Heuristic algorithms Vectors Manifolds Computational modeling Electronic mail Forestry Analytical models

来源：评论

学校读者我要写书评

暂无评论

Centroid Integer Selection Model -- A High Efficiency Method on Dynamic Multi-document Summarization

Centroid Integer Selection Model -- A High Efficiency Method...

引用

International Conference on Asian Language Processing (IALP)

作者： Meiling Liu Dequan Zheng Tiejun Zhao Yang Yu Machine Intelligence & Translation Laboratory Harbin Institute of Technology Harbin China Department of Computer Science and application Northeast Forestry University Harbin China

This paper researches centroid integer selection based on dynamic multi-document summarization (DMS) and presentes a dynamic multi-document summarization model, called Centroid Integer Selection Model (CISM). This model has mainly two steps. First, some abstracts were extracted from the document sets based on different first sentence, respectively. Second, the best abstract was selected based on centroid strategy from all the abstracts created in the first step. The best advantage this model showed was that it eliminated the effect caused by falsely selecting based on the first sentence. Some experiments were conducted on the Update Task test data from TAC2008, and results of new model were compared with results from the TAC2008 evaluation.

关键词： Abstracts Information retrieval Electronic mail Laboratories Silicon Computational modeling Current measurement

来源：评论

学校读者我要写书评

暂无评论

AMT-PSO: An adaptive magnification transformation based particle swarm optimizer

引用

IEICE Transactions on Information and Systems 2011年第4期E94-D卷 786-797页

作者： Zhang, Junqi Ni, Lina Xie, Chen Tan, Ying Tang, Zheng Department of Computer Science and Technology Tongji University Ministry of Education Shanghai 200092 China College of Info Sci and Engi Shandong University of Science and Technology Qingdao China Key Laboratory of Machine Perception Department of Machine Intelligence School of Electronics Engineering and Computer Science Beijing 100871 China Department of Intellectual Information Systems Engineering University of Toyama Toyama-shi 930-8555 Japan

This paper presents an adaptive magnification transformation based particle swarm optimizer (AMT-PSO) that provides an adaptive search strategy for each particle along the search process. Magnification transformation is a simple but very powerful mechanism, which is inspired by using a convex lens to see things much clearer. The essence of this transformation is to set a magnifier around an area we are interested in, so that we could inspect the area of interest more carefully and precisely. An evolutionary factor, which utilizes the information of population distribution in particle swarm, is used as an index to adaptively tune the magnification scale factor for each particle in each dimension. Furthermore, a perturbation-based elitist learning strategy is utilized to help the swarm's best particle to escape the local optimum and explore the potential better space. The AMT-PSO is evaluated on 15 unimodal and multimodal benchmark functions. The effects of the adaptive magnification transformation mechanism and the elitist learning strategy in AMT-PSO are studied. Results show that the adaptive magnification transformation mechanism provides the main contribution to the proposed AMT-PSO in terms of convergence speed and solution accuracy on four categories of benchmark test functions. © 2011 The Institute of Electronics, Information and Communication Engineers.

关键词： Particle swarm optimization (PSO)

来源：评论

学校读者我要写书评

暂无评论

Enhance Top-down method with Meta-Classification for Very Large-scale Hierarchical Classification 5

Enhance Top-down method with Meta-Classification for Very La...

引用

5th International Joint Conference on Natural Language Processing, IJCNLP 2011

作者： Wang, Xiao-Lin Zhao, Hai Lu, Bao-Liang Center for Brain-Like Computing and Machine Intelligence Department of Computer Science and Engineering Shanghai Jiao Tong University 800 Dong Chuan Rd. Shanghai 200240 China MOE-Microsoft Key Laboratory for Intelligent Computing and Intelligent Systems Shanghai Jiao Tong University 800 Dong Chuan Rd. Shanghai 200240 China

ISBN: (纸本)9789744665645

Recent large-scale hierarchical classification tasks typically have tens of thousands of classes as well as a large number of samples, for which the dominant solution is the top-down method due to computational complexity. However, the top-down method suffers from accuracy deficiency, that is, its accuracy is generally lower than that of the flat approach of 1-vs-Rest. In this paper, we employ meta-classification technique to enhance the classifying procedure of the top-down method. We analyze the proposed method on the aspect of accuracy, and then test it with two real-world large-scale data sets. Our method both maintains the efficiency of the conventional top-down method and provides competitive classification accuracies. © 2011 AFNLP

关键词： Natural language processing systems

来源：评论

学校读者我要写书评

暂无评论

An Empirical Comparative Study on Two Large-Scale Hierarchical Text Classification Approaches

引用

International Journal of computer Processing of Languages 2011年第4期23卷 309-325页

作者： JIAN ZHANG HAI ZHAO LIQING ZHANG BAO-LIANG LU Center for Brain-Like Computing and Machine Intelligence Department of Computer Science and Engineering Shanghai Jiaotong University Shanghai 200240 China MOE-Microsoft Key Laboratory for Intelligent Computing and Intelligent Systems Shanghai Jiaotong University 800 Dongchuan Rd. Shanghai 200240 China Department of Computer Science Virginia Tech Blacksburg VA 24061 USA

Patent classification is a large scale hierarchical text classification (LSHTC) task. Though comprehensive comparisons, either learning algorithms or feature selection strategies, have been fully made in the text categorization field, few work was done for a LSHTC task due to high computational cost and complicated structural label characteristics. For the first time, this paper compares two popular learning frameworks, namely, hierarchical support vector machine (SVM) and k -nearest neighbor ( k -NN) that are applied to a LSHTC task. Our experimental results show that the latter outperforms the former for the LSHTC task, which is quite different from the existing results for normal text categorization tasks. In addition, this paper compares different similarity measures and ranking strategies in k -NN framework for LSHTC task. From our empirical study, conclusions can be drawn that k -NN is more appropriate for the LSHTC task than hierarchical SVM. BM25 outperforms other similarity measures and ListWeak gains a better performance than other ranking strategies. Our empirical results also indicate that using all the labels of the retrieved neighbors can remarkably improve classification performance over only using the first label of the retrieved neighbors.

关键词： Large-scale text classification hierarchical text classification text classification k -NN hierarchical SVM similarity measure ranking strategies

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：