检索结果-内蒙古大学图书馆

A feature combination method for semantic role classification

Journal of Information and Computational Science 2010年第1期7卷 127-133页

作者： Li, Shiqi Zhao, Tiejun Li, Hanjing Liu, Shui MOE-MS Key Laboratory of Natural Language Processing and Speech Harbin Institute of Technology Harbin 150001 China

This paper proposes a Chinese semantic role classification approach on the basis of feature combination. First we define a set of effective basic features. Then a statistics-based feature combination method is developed for constructing a combined feature set from the basic feature set. According to the distribution of each combined feature in positive and negative instances, we utilize a new statistics to measure the classifying performance of the feature, and then put the features with high ratio in the combined feature set. Finally, the semantic roles are classified by SVM classifier using both basic and combined feature set. The experimental results on Chinese Proposition Bank show that the method can improve the performance of SRC by two percent. Copyright © 2010 Binary Information Press.

关键词： Support vector machines

来源：评论

学校读者我要写书评

暂无评论

A clustering based fast detection algorithm for large scale duplicate emails

A clustering based fast detection algorithm for large scale ...

引用

International Conference on Machine Learning and Cybernetics (ICMLC)

作者： Lin Sun Bing-Quan Liu Bao-Xun Wang Xiao-Long Wang MOE-MS Key Laboratory of Natural Language Processing and Speech Harbin Institute of Technology Harbin China

Duplicate emails, which exist on the internet widely and are mainly caused by mailing lists, not only waste storage resource but also bring users garbage. In this paper, according to the structure and text feature of email, we put forward the concept of Mail-Duplicate-Degree, and in this way the email duplicate is firstly defined. Based on this definition, we develop an algorithm based on clustering to detect duplicate emails. By introducing a hash function provided by TRIE tree to optimize the efficiency, the algorithm gets over the slow processing speed problem existing in traditional clustering methods. Experimental results on large-scale emails have shown that the algorithm has a high precision.

关键词： Electronic mail Clustering algorithms Internet Noise Algorithm design and analysis Feature extraction Layout

来源：评论

学校读者我要写书评

暂无评论

Event Entailment Extraction Based on EM Iteration

Event Entailment Extraction Based on EM Iteration

引用

International Conference on Asian language processing (IALP)

作者： Zhen Li Hanjing Li Mo Yu Tiejun Zhao Sheng Li MOE-MS Key Laboratory of Natural Language Processing and Speech Harbin Institute of Technology Harbin China

In the research and development of various natural language processing systems, like Q&A system and text-to-scene conversation system, we realize that knowledge of text entailment helps a lot in improving the performance of the system. Systems with text entailment knowledge will be smarter than those who without entailment knowledge. Currently many research teams are focusing on text entailment, including recognition, generation and extraction. However, entailment extraction is the main method in creating entailment knowledge database. Meanwhile, for text-to-scene conversation system, due to the importance of events in stories, find a method for event entailment extraction is our main goal. In this paper, we proposed a method for extracting event entailment from corpus based on EM iteration, which has not been used before.

关键词： Context Internet Animation natural language processing Libraries Classification algorithms Encyclopedias

来源：评论

学校读者我要写书评

暂无评论

Study on the optimal parameters of image fusion based on wavelet transform

引用

Journal of Computational Information Systems 2010年第1期6卷 131-137页

作者： Zheng, Hong Zheng, Dequan Hu, Yanxiang Li, Sheng MOE-MS Key Laboratory of Natural Language Processing and Speech Harbin Institute of Technology Harbin 150001 China Tianjin Normal University Tianjin 300387 China

Selection of wavelet type, decomposition level and fusing rule is a key problem when wavelet transform is applied to image fusion. 2916 kinds of different fusing methods(54×5×9, including 54 wavelet types, 5 decomposing levels, and 9 fusing rules)are analyzed and compared in the experiment of fusing multi-focus images in this paper. Through calculating the comparability degree of fused images, the fusion performances are evaluated. And the experiment shows that the similarities of the results and the ideal pictures are all over 0.999, showing pretty good performance. Copyright © 2010 Binary Information Press.

关键词： Image fusion

来源：评论

学校读者我要写书评

暂无评论

An information extraction system for heterogeneous Web source

An information extraction system for heterogeneous Web sourc...

引用

International Conference on Machine Learning and Cybernetics (ICMLC)

作者： Ting Zhou Cheng-Jie Sun Lei Lin Bing-Quan Liu MOE-MS Key Laboratory of Natural Language Processing and Speech School of Computer Science and Technology Harbin Institute of Technology Harbin China

Information Extraction is the task of identifying information in texts and converting it into a predefined format. In this paper, we build an information integration system which focuses on the information of computer science teachers in Chinese universities. The target of the system is to automatically extract the useful information from heterogeneous sources and re-organize them into structured format. The system includes 4 main modules: web pages retrieval module, web pages' structure classification module, information extraction module and information updating module. We have successfully applied the system to deal with 107 universities in China which shows the effect of the proposed system.

关键词： Web pages Data mining Classification algorithms Educational institutions Crawlers Support vector machines Search engines

来源：评论

学校读者我要写书评

暂无评论

Research on Domain-Adaptive Transfer Learning Method and Its Applications

Research on Domain-Adaptive Transfer Learning Method and Its...

引用

International Conference on Asian language processing (IALP)

作者： Geli Fei Dequan Zheng MOE-Microsoft Key Laboratory of Natural Language Processing and Speech Harbin Institute of Technology Harbin China

Traditional machine learning methods rely on strong assumptions, especially assuming that training data and testing data in homogeneous feature spaces. However, this is not always true in reality. To break such assumptions, this paper proposes a domain-adaptive transfer learning method, which automatically learns knowledge from existing knowledge bank by extracting linguistic information such as part-of-speech and co-occurrence of keywords and constructing a new domain-adaptive transfer knowledge bank. Through experiments on homogeneous and heterogeneous feature spaces, we testify the efficacy of our methods.

关键词： Text categorization Learning systems Art Training data Testing Machine learning Feature extraction

来源：评论

学校读者我要写书评

暂无评论

Event entailment chains extraction for Text-to-Scene conversion

Event entailment chains extraction for Text-to-Scene convers...

引用

International Conference on Machine Learning and Cybernetics (ICMLC)

作者： Han-Jing Li Zhen Li Xiao-Ping Xue Tie-Jun Zhao MOE-MS Key Laboratory of Natural Language Processing and Speech Harbin Institute of Technology Harbin China Department of mathematics Harbin Institute of Technology Harbin China

In our study of Text-to-Scene conversation (TTS), which translates natural language into animations automatically, we realized that event entailment knowledge is useful in generating scenes since the main part of a scene is to show an event. In this paper, we provide some results of our attempt to extract event entailment knowledge. We use entailment chains instead of traditional entailment rules since the sequence of events is a process which make useful in TTS. The result shows that the work is worth to continue to study.

关键词： Animation Machine learning Cybernetics Joining processes Generators natural language processing

来源：评论

学校读者我要写书评

暂无评论

A study of features on Primary Question detection in Chinese online forums

A study of features on Primary Question detection in Chinese...

引用

International Conference on Fuzzy Systems and Knowledge Discovery (FSKD)

作者： Lin Sun Bingquan Liu Baoxun Wang Deyuan Zhang Xiaolong Wang MOEMS Key Laboratory of Natural Language Processing and Speech Harbin Institute of Technology Harbin China

Primary Question detection in online forum is a subtask of extracting question-answer pairs. In this paper, by surveying the forms of questions in Chinese online forums, a combination of textual and N-gram features achieved via feature selection is adopted to help detecting primary questions. By viewing primary question detection a binary classification problem, decision tree classifier C4.5 and support vector machine are introduced to distinguish questions from non-questions separately. Experimental results across multiple datasets demonstrate that the mixture of textual and N-gram features performs better than using each of them separately under both C4.5 and support vector machine. By computing the weight of each feature in the two classifiers, the top 6 features are found the very same except for a little adjustment of order, showing that the combination of textual and N-gram features is universal and effective in detecting primary questions.

关键词： Feature extraction Support vector machines Electronic mail Classification tree analysis speech

来源：评论

学校读者我要写书评

暂无评论

Semi-supervised domain adaptation for WSD: Using a word-by-word model selection approach

Semi-supervised domain adaptation for WSD: Using a word-by-w...

引用

IEEE International Conference on Cognitive Informatics

作者： Yuhang Guo Wanxiang Che Ting Liu Sheng Li Research Center of Information Retrieval MOE-Microsoft Key Laboratory of Natural Language Processing and Speech School of Computer Science and Technology Harbin Institute of Technology China

This paper proposes a word-by-word model selection approach to domain adaptation for Word Sense Disambiguation. By this approach, the model for a target word is automatically selected from a candidate model set, which is comprised of improved self-training models and a supervised model. The improved self-training uses sense priors to prevent its iteration from converging into undesirable states. Experimental results on a domain-specific corpus show that: (1) our improved self-training model is effective for the words which have target domain linked senses; (2) the selected models obtain higher accuracies than each single model and effectively improve the performance compared to the state-of-the-art supervised model.

关键词： Support vector machines Adaptation model Training Accuracy Finance Data models Training data

来源：评论

学校读者我要写书评

暂无评论

AN EFFICIENT APPROACH TO COMMENT SPAM IDENTIFICATION

引用

Journal of Electronics(China) 2009年第5期26卷 644-650页

作者： Yang Yuhang Zhao Tiejun Zheng Dequan Yu Hao MOE-MS Key Laboratory of Natural Language Processing and Speech Harbin Institute of Technology Harbin 150001 China

This paper proposes a novel approach to comment spam identification based on content analysis. Three main features including the number of links, content repetitiveness, and text similarity are used for comment spam identification. In practice, content repetitiveness is determined by the length and frequency of the longest common substring. Furthermore, text similarity is calculated using vector space model. The precisions of preliminary experiments on comment spam identification conducted on Chinese and English are as high as 93% and 82% respectively. The results show the validity and language independency of this approach. Compared with conventional spam filtering approaches, our method requires no training, no rule sets and no link relationships. The proposed approach can also deal with new comments as well as existing comments.

关键词： Comment spam Automatic identification Content analysis Blog

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：