检索结果-内蒙古大学图书馆

New approach to query expansion in information retrieval

High technology Letters 2008年第1期14卷 77-80页

作者： Li, Weijiang Zhao, Tiejun Wang, Xian'gang Key Laboratory of Natural Language Processing and Speech School of Computer Science and Technology Harbin Institute of Technology Harbin 150001 China

To eliminate the mismatch between words of relevant documents and user's query and more serious negative effects it has on the performance of information retrieval, a method of query expansion on the basis of new terms co-occurrence representation was put forward by analyzing the process of producing query. The expansion terms were selected according to their correlation to the whole query. At the same time, the position information between terms were considered. The experimental result on test retrieval conference (TREC) data collection shows that the method proposed in the paper has made an improvement of 5%-19% all the time than the language modeling method without expansion. Compared to the popular approach of query expansion, pseudo feedback, the precision of the proposed method is competitive.

关键词： Information retrieval

来源：评论

学校读者我要写书评

暂无评论

引用

Journal of Computational Information Systems 2008年第2期4卷 739-746页

作者： Yu, Feng Zheng, Dequan Zhao, Tiejun Li, Sheng Yu, Hao School of Computer and Information Engineering Harbin University of Commerce Harbin 150076 China Key Laboratory of Natural Language Processing and Speech Harbin Institute of Technology Harbin 150001 China

A novel solution of computing document similarity based on a combination of ontology with statistical method was presented in this paper. In this study, word sense ontology of keyword is firstly constructed by context multi-information, and then, the same feature firstly was acquired from document pairs, the usage of context co-occurrence feature was gotten in using part of speech, semantic, location, average co-occurrence probability, and was expressed as the linguistic ontology knowledge;final, document similarity evaluation value is calculated for each document to judge the document similarity degree. The Chinese document set from the NTCIR-3 workshop collection was used to evaluate the method, an average precision increase can be achieved at top 10 and 100 ranking.

关键词： Text processing

来源：评论

学校读者我要写书评

暂无评论

Japanese NER Post-processing Based on Improved TBL Method

Japanese NER Post-Processing Based on Improved TBL Method

引用

International Conference on Fuzzy Systems and Knowledge Discovery (FSKD)

作者： Dequan Zheng Jing Wang Tiejun Zhao MOE-MS Key Laboratory of Natural Language Processing and Speech Harbin Institute of Technology

A TBL based post-processing approach is proposed for Japanese named entity recognition (NER) in this paper. Firstly, tuning rules are automatically acquired from the results of Japanese NER by error-driven learning. And then, the tuning rules are optimized according to given threshold conditions. After filtered, the rules are used to revise the results of Japanese NER. Above all, this approach could be used in special domains perfectly for its learning domain linguistic knowledge automatically. The learnt rules could not go over fit as well. The experimental results show that a high result can be achieved in precision for Japanese NER.

关键词： Hidden Markov models Degradation Error correction Fuzzy systems Laboratories natural language processing speech processing Graphics speech recognition Automatic speech recognition

来源：评论

学校读者我要写书评

暂无评论

Identifying named entities in biomedical text based on stacked generalization

Identifying named entities in biomedical text based on stack...

引用

World Congress on Intelligent Control and Automation (WCICA)

作者： Haochang Wang Tiejun Zhao MS Key Laboratory of Natural Language Processing and Speech in Harbin Institute of Technology MOE Harbin China

Biomedical named entity recognition is a basic technique in the biomedical knowledge discovery and its performance has direct effects on further discovery and processing in biomedical texts. In this paper, we present stacked generalization strategy for biomedical named entity recognition including homogeneous classifier ensembles and heterogeneous classifier ensembles based on stacked generalization. Evaluations show that stacked generalization strategy can take advantage of more useful evidences, and make use of compensation and relativity among different classifiers to learn the correlation between individual classifiers predictions and the correct prediction to improve the performances of the system. This method breaks through the limitation of single classifier and achieves promising performances.

关键词： Classification algorithms Training Data mining Stacking Data models Support vector machine classification Training data

来源：评论

学校读者我要写书评

暂无评论

A Fusion of Multiple Classifiers Approach Based on Reliability function for Text Categorization

A Fusion of Multiple Classifiers Approach Based on Reliabili...

引用

International Conference on Fuzzy Systems and Knowledge Discovery (FSKD)

作者： Qingxuan Chen Dequan Zheng Tiejun Zhao Sheng Li MOE-MS Key Laboratory of Natural Language Processing and Speech Harbin Institute of Technology Harbin China

With the development of Internet and the rapid expansion of electronic resource, text classification technology is becoming an effective organization and management tool to deal with information. In this paper, a method for text categorization based on the fusion of multiple classifiers was presented, reliability function was introduction to select the text that hard to give determine by the main classifier, for these texts, multiple classifiers were used to give the determine which category the unlabeled documents belong to by voting. Experiments showed that the performance of text classification improved by the proposed method. Compared with single classifier, this method achieved better performance, only increasing a small amount of time than using single main classifier. Besides this, this method is more stable than using single classifier for text categorization task, especially when using different corpuses to check the performance of various methods.

关键词： Text categorization Voting Support vector machines Support vector machine classification Frequency shift keying Fuzzy systems Laboratories natural language processing speech processing Internet

来源：评论

学校读者我要写书评

暂无评论

Query Translation Selection for Cross-language Information Retrieval Based on HowNet

Query Translation Selection for Cross-Language Information R...

引用

The 2008 IEEE International Conference on natural language processing and Knowledge Engineering(IEEE NLP-KE 2008)(2008IEEE自然语言处理与知识工程国际会议)

作者： Honglei ZHU Dequan ZHENG Tiejun ZHAO MOE-MS Key Laboratory of Natural Language Processing and Speech Harbin Institute of TechnologyHarbin 150001

Research on cross-language information retrieval (CLIR) increasingly concentrates in candidate translation selection of the keywords in the query. The accuracy of translation has a direct impact on accurate rate and recalled rate. This thesis presents three methods based on HowNet to resolve query translation ambiguity of CLIR. The first is based on semantic relation, and it uses semantic relation network of context to determine the semantic of keywords and then select the correct translation. Bilingual decaying co-occurrence model count bilingual corpus co-occurrence information which includes the times and distance value of co-occurrence, which is different from monolingual co-occurrence. To resolve the problem of sparseness in corpus and make full use of the bilingual corpus, this paper gives another model that is semantic decaying co-occurrence model. Through test and summarizing this paper gets the best algorithm to integrate the traits of the three models, which gradually optimizes the translation and gets a higher precision.

关键词： Query translation CLIR statistical method translation selection OOV

来源：评论

学校读者我要写书评

暂无评论

Research on Query Translation Disambiguation for CLIR Based on HowNet

Research on Query Translation Disambiguation for CLIR Based ...

引用

The 9th International Conference for Young Computer Scientists(第九届国际青年计算机大会)

作者： Honglei Zhu Dequan Zheng Tiejun Zhao MOE-MS Key Laboratory of Natural Language Processing and Speech Harbin Institute of Technology Harbin 150001

Query translation is an important task for crosslanguage information retrieval (CLIR), which aims at translating the query described in source language into target language. The approach to query translation based on bilingual dictionary is becoming the mainstream thinking because of its simplicity and the increasing availability of machine readable bilingual dictionary. However, this kind of approach faces two necessary problems that is ambiguity in translation and the incompleteness of the dictionary. This paper focuses on the first problem, and it presents three statistical models based on HowNet to resolve query translation ambiguity of CLIR: Query translation selection based on semantic relation;Bilingual decaying co-occurrence model and Semantic decaying co-occurrence model. Through test and summarizing this paper gets the best algorithm to integrate the traits of the three models, which gradually filters and optimizes the translation.

关键词： Query translation CLIR statistical method translation selection OOV.

来源：评论

学校读者我要写书评

暂无评论

A new approach to query expansion in information retrieval

引用

High technology Letters 2008年第1期14卷 77-80页

作者：李卫疆 Zhao Tiejun Wang Xian＇gang MOE-MS Key Laboratory of Natural Language Processing and Speech -School of Computer Science and Technology Harbin Institute of Technology Harbin 150001 P.R. China

To eliminate the mismatch between words of relevant documents and user's query and more seriousnegative effects it has on the performance of information retrieval,a method of query expansion on the ba-sis of new terms co-occurrence representation was put forward by analyzing the process of *** expansion terms were selected according to their correlation to the whole *** the sametime,the position information between terms were *** experimental result on test retrievalconference(TREC)data collection shows that the method proposed in the paper has made an improve-ment of 5%～19% all the time than the language modeling method without *** to thepopular approach of query expansion,pseudo feedback,the precision of the proposed method is competi-tive.

关键词： information retrieval language model query expansion

来源：评论

学校读者我要写书评

暂无评论

Structure and Content Based Blog Pages Identification

Structure and Content Based Blog Pages Identification

引用

International Conference on Fuzzy Systems and Knowledge Discovery (FSKD)

作者： Feng Yu Dequan Zheng Tiejun Zhao Xiao Cheng School of Computer and Information Engineering Harbin University of Commerce China MOE-MS Key Laboratory of Natural Language Processing and Speech Harbin Institute of Technology

Blog is becoming more and more popular with the rapid development of Internet. It needs to find an automatic way to distinguish the blog pages from ordinary Web pages for the content extraction of blog pages and the blog community discovered. Some basic concepts and ideas in the area of blog was described in this paper, and a method on the blog pages identification is proposed, which is based on the blog pages structure and blog content. The experimentation shows that a high result can be achieved in precision.

关键词： Information services Web sites Internet Web pages Navigation Support vector machines Support vector machine classification Fuzzy systems Knowledge engineering Business

来源：评论

学校读者我要写书评

暂无评论

ORDINAL REGRESSION FOR INFORMATION RETRIEVAL

引用

Journal of Electronics(China) 2008年第1期25卷 120-124页

作者： Qi Haoliang Li Sheng Gao Jianfeng Han Zhongyuan Xia Xinsong Ministry of Education-Microsoft Key Laboratory of Natural Language Processing and Speech Harbin Institute of Technology Harbin 150001 China Microsoft Research One-Microsoft Way Redmond WA 98052 USA Heilongjiang Institute of Technology Harbin 150001 China Peking University Beijing 100871 China

This letter presents a new discriminative model for Information Retrieval (IR), referred to as Ordinal Regression Model (ORM). ORM is different from most existing models in that it views IR as ordinal regression problem (i.e. ranking problem) instead of binary classification. It is noted that the task of IR is to rank documents according to the user information needed, so IR can be viewed as ordinal regression problem. Two parameter learning algorithms for ORM are presented. One is a perceptron-based algorithm. The other is the ranking Support Vector Machine (SVM). The effec- tiveness of the proposed approach has been evaluated on the task of ad hoc retrieval using three English Text REtrieval Conference (TREC) sets and two Chinese TREC sets. Results show that ORM sig- nificantly outperforms the state-of-the-art language model approaches and OKAPI system in all test sets; and it is more appropriate to view IR as ordinal regression other than binary classification.

关键词： Information Retrieval （IR） Ordinal Regression Perceptron Ranking Support Vector Machine （SVM）

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：