This paper reviews iPRES2009 digital preservation international conference comprehensively, focusing on “Moving Into the Mainstream, Enabling Our Digital Future”, analyzes and discusses preservation infrastructure、research data & workflows、sustainability & cost models、metadata & important property、formats、preservation practice and case studies deelply. In particular,it stresses that the promise of digital preservation will be realized when it is truly integrated into the mainstream of digital scholarship, culture, and commerce.
By comparing with extracting bilingual terminology from parallel corpora, this paper describes the value of extracting bilingual terminology from comparable corpora. It summarizes the main method and the optimization methods of implementation of bilingual terminology extraction. And some perspectives and prospects about bilingual terminology extraction based on the comparable corpus are proposed.
The paper summarizes the current research on the concept of context. Besides, it introduces the levels of context and makes a review on this field from both of research on theories and empirical studies, including information environment level, information seeking level, information retrieval interaction level and query level. Additionally, it outlines the research existing problems and the future directions of context in information retrieval.
The paper analyzes the research and implementation algorithm about identifying the maximal meaningful node. Making uses of and improving the style tree，it computes the importance of nodes to find the maximal meaningful node. Finally， an example is given.
This paper introduces several technologies to realize Single Sign On system of library，which can deal with user inconvenience and stress of repeated login. After comparing advantages and disadvantages of commercial software and open source software, CAS is chosen for the Single Sign On system. According to the actuality of service system of the National Science Library.It expounds processes of realizing the system. The design makes it more convenient when the librarian and user use the different service system, and can provide a reference to make up relative Single Sign On system of library.
To help e-Commerce websites provide personalized recommendation management based on collaborative filtering, an e-Commerce collaborative filtering prototype that is called ECRec, is proposed and implemented. ECRec includes two basic algorithms and four improved algorithms, and its architecture is independent on e-Commerce business systems，consequently, ECRec has a better portability and maintainability. Moreover, the algorithm interface in ECRec is embedded, thus ECRec has the characteristics of open architecture, and websites can add more collaborative filtering algorithms into ECRec.
Due to the hardware limitations of mobile terminal equipment and keywords submited by users,there are problems of word mismatch between short queries and query results. A mobile query expansion method based on related words co-occurrence strategy is proposed, which is called ALRCO.It utilizes the related words co-occurrence information in the abstract of documents and keywords in the query logs to evaluate quality of the expansion words, and selects the most appropriate expansion terms. The expansion words with the initial query have the better relevance to the characterization of the theme.Finally,experimental results show that ALRCO offers more accuracy compared with traditional query.
This paper summarizes the current situation of Web Performance Testing Models（ WPTM ）and the accessing characteristic of users, and proposes a new WPTM based on Web information system.It increases integrative performance indicator which includes actual require time and requirement success ratio to aid testing, and improves the testing process. Finally，an instance is given to verify the correctness and effectiveness.
Concerning the present problem of a growing academic plagiarism，the algorithm of the text copy detection based on text structure tree is put forward．A paper can be divided into a construction tree with three layers：the uppermost root node is a text；branch node represents a sentence bag；leaf node denotes sentence.According to synthetic similarity and a function this paper computes sentence similarity，and similarity of leaf node is based on maximal sentence similarity．At the same time，the upper similarity is derived from the adjacent lower similarity．Finally，papers of China Journal Full-Text Database is chosen for a test，and the experimental result shows that this algorithm is feasible and efficient．
Based on the keywords in the 11 261 papers in the field of information services from CNKI, this paper constructs an undirected weighting network which contains 6 401 vertices(keywords) and 21 007 edges using co-word analysis, and verifies that the network has the characters of scale free and small world. The index of degree centrality and betweenness centrality of vertices in the network are calculated, and a method of detecting cross concept in the network is introduced. Finally, using the G-N clustering algorithm, the paper performs a cluster analysis on the domestic information services research concept network, and divides the research field into 7 different branches.
This paper summarizes the laboratory information characters based on analysis of university laboratory Web information, which is used to formulate rules of laboratory Web information.It designs an information extraction system on university laboratory, and presents system architecture and technical architecture of labIE. It also describes the design of rules on table recognition and methodology of constructing characteristic predicate.
To deal with “disaster of dimensionality”, cluster identifying and large-scale problems arising in text clustering algorithm’s applications, a parallel text clustering method is proposed and implemented,which uses WordNet to the dimensionality reduction of the word list and stemming based on POS tagging and WordNet. Comparing with the Porter Stemming method, the experimental results show that this method can substantially reduce the dimension of word list, improve the accuracy and recall rate of the clustering and have a better understanding of each cluster.
This paper introduces the work experience of Foreign Teaching Materials Center of Tsinghua University Library on the construction of the cooperate building and resource sharing information system taking the Foreign Teaching Materials Center of China Education Ministry Information System as an example. The platform construction is described in detail from the aspects of data norm, system design, function realization，platform feature and so on.
Concerning on various stages of periodicals management including current issues ordering, acceptance and binding, the paper converts and combines available external data and internal data of ILAS automation system from the data selection, ILAS format requirements, data conversion, data access and quality control. On the basis of that, it achieves the various types of data processing and automation in periodicals management, and puts forward higher requirements for periodical management automation system.
Short message service is a new method of information service in digital library. The paper designs and implements an interactive short message service platform based on LIBSYS, and solves some most important problems such as creating，sending，receiving and dealing with short messages.
According to the practical requirement of document management in National Engineering Technology Library (NETL), this paper designs a standing order management system. It analyses the characteristics of standing order in terms of acquisition, listing and price, discusses the solution to these characters, and designs the operation flow of order administration, documents (acceptance/listing) administration and settlements administration. The system is overall integrated with MELINETS from data to functions, therefore it realizes the automatic management of standing orders.
The paper firstly introduces the design ideas of the Academic Papers Management System and the methods of data collection.Then it describes the methods of constructing the Academic Papers Management System based on DSpace. Finally， the application effect and future development of the system are discussed..
