Personalized service is a way to optimiz information system.This paper introduces the models of user behaviour to optimize retrieval systems, information recommendation systems, workflow management systems, user-generated system, social network systems, media player system, Web navigation support system, mobile information system, and interactive panels. The change of user models are explained, from model of using tools, to model of technical operations, then to model of user psychology and behavior.To propose a workflow including “User Behavior-User Modeling-Personalized Services-Redesign” as the overall program that user behaviour driven digital library personalized services.
This paper tries to change the traditional analysis mode that using association rule mining to gain the subject relationship based on single standard, and introduces the Ontology mechanism with semantic description capabilities into the knowledge organization of CSSCI academic resource for organizing subject and related concepts by object-oriented approach, so that to establishes CSSCI academic resource networks model based on Ontology. Then subject evaluation method is used to analyze the relationship between subjects annotated in CSSCI_Onto, and knowledge mining technique is also adopted to discover the multi-subject association patterns that users are interested in and implies in original knowledge, by which to obtain analysis conclusion for supporting decision, and to provide factual basis for interdisciplinary cooperation enhancement and cross-disciplinary, frontier-disciplinary emergence and development.
This paper firstly builds concept lattice of some ball-games with ConExp1.3 and Lattice Miner1.4. Then it compares the quality and operation of the two tools from the basic information, modification of formal context, layout of lattice, mining of association rules and storage management. ConExp stresses the concept and the relationships of concepts, and personalized presentation of the concept lattice; and Lattice Miner has advantages to deal with the complex problem, extract association rules, and support semantic network. It makes the foundation for the research based on concept lattice tool.
This paper analyzes the contradiction between the needs of catalog development and the limitations of existing catalog systems, then studies Service-Oriented Architecture(SOA) and constructs a catalog Model by Service-oriented method based on SOA. It creates catalog Micro-Services, which are loose-coupling, autonomic and composable in code layer. The SOA model establishes catalog on demand by Micro-Services restructuring mechanism, and provides system support for catalog development. It has a high degree of openness and scalability,which can promote the development of catalog,and can be used as references for other library services.
In order to solve the problem of fuzzy semantic in social annotation system, this paper analyses the relation among users, resources and tags, introduces latent semantic analysis probabilistic PLSA model. By extending PLSA model,the annotation is mapped to a finite-dimensional latent semantic space, and the collection of latent semantic of annotation is obtained by clustering. This discovery method improves the satisfaction of user’s actual need for resource in social annotation system. Finally, experimental results show the effectiveness of the proposed method.
This paper demonstrates the concept, structure and functions, and characteristics of SKOS. And also concretely analyzes the realization of SKOS’ systematic and standardized control of the Chinese Archival Thesaurus. Finally it summarizes its characteristics such as standardization, systematization, flexibility, and practicability,etc.
According to extraction of hot keywords in the multi-phase candidate keywords, the paper tries mass data process,determines the meaningless words based on the timing of statistical law, and proposes Union Variance (UV) concept. The HK (Hot Keywords) formula is constructed based on multi-feature fusion to achieve the extraction of hot keywords. Experimental results show that this method is efficient in the process of hot subject extraction.
This paper suggests that trust is another important factor effecting recommendation result and introduces trust- worthiness into traditional collaborative filtering algorithm. It proposes a collaborative filtering recommendation algorithm based on improved trustworthiness,which combines similarity and trustworthiness to substitute traditional similarity weight. The experiment results can prove the validity and superiority of the proposed algorithm.
This paper introduces a Web page extraction algorithm named WEAV(Web-page Extraction Algorithm based on VIPS).WEAV is used in a mobile meta search engine named M-Meta which is designed for extracting the main content of Web pages and returning them to users. And it makes the result be adaptive for mobile devices displaying, improves the retrieval speed and strengthens the usability of Web on mobile devices.
A method based on the Conditional Random Fields (CRFs) is proposed to extract the information of unstructured factual information text, and the method of parameter estimation and feature selection is also anlyzed. During information extraction, the author blocks the text firstly with the help of format information such as separator and special identifier, and then extracts the designated block with Conditional Random Fields. The proposed method is applied in Global Weapon Knowledge Base System (GWKBS), and experiment results show that it has a better precision and recall performance.
The paper focuses on a large number of news corpus, pretreats the titles and abstracts of training documents, then builds up the feature vector library. At last, it uses matching method of decision table rules and vector space method to identificate the articles in two ways, and makes better service of the sudden events recognition on Web.
This article makes an in-depth Web content analysis about the essential attributes of Web usability on basis of the whole Internet data acquired by some tools, then further clarifies the conception of Web usability and constructs the logical levels of its evaluation index system. The author carries out Web usability evaluation on the portals of 29 representative universities in Jiangsu Province by using Web engine and related software.
Aiming to meet the internal data processing needs of information organizations, this paper, by analyzing the frameworks of Amazon Elastic Map/Reduce (EMR) platform, puts forward to build the dynamic and elastic open source mass data mining platform based on cloud computing, and provides a roadmap of successful implementation, an example of massive text data processing and the analysis of advantages of open source EMR platform. This implementation plan includes three parts: building dynamic virtual environment of cloud computing,creating the virtual server template of Hadoop, and deploying and running Cloudera and Cloudera Desktop. Through the application of open source EMR platform , the problem of server sprawl can be solved effectively,the utilization ratio of network computing resource is improved,and the rapid deployment capability and agility of distributed data processing services are enhanced.
Integrated retrieval mechanism is studied for open access system and the Web crawling is used to build a distributed DSearch system based on Nutch, which can provide a kind of efficient, flexible, customizable search tools. Three key technologies are also introduced,including distributed cluster configuration,Chinese word splitter modification and index settings. Finally,the functions of DSearch are evaluated with the selected feed lists.
The paper introduces the concept of Mashup and its basic application. With the analysis of Douban.com, it combines the Douban’s books appraisal recommendation function and the library OPAC system by using Mashup technology, and enhances the library service ability. The implementation idea and the essential code are given,as well as the practice of combination with Douban and OPAC of the Nanjing University Library. The achievement obtains users’ approval and preference.
Through analyzing OAI XML and the TRS system metadata format of the dissertation, referring to the relevant standards and norms of CALIS, the article proposes a method with VB programming, Which can achieve and harvest the OAI XML metadata of dissertation by dissertation central database. As a result, the method provides a solution to non-standard dissertation systems.
