[Objective] This paper is an in-depth review of popular research methodologies adopted by the Citation-Based Summarization (CBS) studies. [Coverage] We retrieved scholarly papers on CBS published since 2007, as well as earlier research on automatic summarization and citation analysis. [Methods] We thoroughly discussed the basic concepts and natural language processing technology in the field of CBS. [Results] Citances plays more important roles in automatic summarization applications than randomly selected sentences from scientific works. [Limitations] We did not compare the current achievements with possible results under the ideal circumstances. [Conclusions] CBS technology expands the scope of traditional informetrics and automatic summarization studies. It also offers suggestion to improve the existing evaluation methods of automatic summarization services. CBS calls for the expansion of citation windows and new experimental corpus. We have addressed these issues and explored new perspectives for the CBS research.

[Objective] We try to predict the future trends of Meta-analysis methodology, and improve the performance of domain knowledge discovery tasks. [Methods] First, we reviewed the features of popular Meta-analysis software, as well as their differences in computing and graphics functions. Second, we designed a Meta-analysis tool for the resources and environment science. [Results] We developed a new concept for public oriented Meta-synthesis tool with standardized interface, simplified procedure and accurate results. [Limitations] We did not examine the feasibility of the new tool on a working platform, The inherent weaknesses of Meta-analysis cannot be avoided in the domain knowledge discovery. [Conclusions] We need to build a platform for the domain knowledge discovery with Meta-analysis technology and then expand its application in literature discovery.

[Objective] This research establishes an assessment mechanism for digital library’s micro-services from the perspectives of user, information and technology. It provides the theoretical foundation and criteria for us to improve the micro-services. [Methods] First, we proposed a Group AHP method based on clustering technique to construct a judgment matrix to evaluate the target system. Second, conducted an empirical analysis of 10 representative digital libraries from China and abroad with the fuzzy membership degree function. [Results] The proposed mechanism was practical and could help us improve the micro-services of digital libraries. [Limitations] The assessment criteria should be further examined under the specific circumstances facing the digital libraries. The sample size of the empirical analysis is small, thus, the conclusion might be limited and subjective. [Conclusions] This paper provides new research perspectives for the digital library micro-services, as well as methods to improve their quality and efficiency.

[Objective] We propose an effective method to cluster and discover the needed Web services. [Methods] First, we employed the Biterm Topic Model to learn the latent topics of the Web service description corpus. Second, we retrieved and clustered each document’s topic distribution. Finally, we created a mechanism to discover Web service quickly. [Results] The proposed method achieved better precision rate and normalized discounted cumulative gain than methods using Latent Dirichlet Allocation and external corpus. [Limitations] Only considered functions of the Web services, and did not include the quality factors to the algorithm. [Conclusions] The proposed method could identify the needed services more accurately.

[Objective] The paper proposes a method to create new Ontology for reusable products based on the concept lattice theory, aiming to improve the traditional methodology. [Methods] First, we extracted classes, attributes and relationships data from product categories and made appropriate adjustments. Second, we created formal contexts and concept lattices to generate product Ontology automatically. [Results] The proposed method improved the product classification scheme and enhanced the reusability and sharing of Ontology. It also visualized the existing and potential entities as well as their relationships. [Limitations] Only used a segment of medical product Ontology to implement the proposed method, and then discussed its flexibility for eCl@ss product classification system. [Conclusions] The field and scope of Ontology need to be clearly defined before we construct product Ontology based on concept lattice. The categories should be sorted and grouped with scientific classification principles. We also need to build the property dictionary and define object properties. The Ontology will be generated automatically with formal context and concept lattice.

[Objective] To analyse the linguistic features of new findings discussed by the scientific research papers in Chinese. [Methods] We first annotated these features and then explore their patterns with the help of sentence analysis, frequency statistics and co-occurrence analysis technologies. [Results] We summarized the sentence patterns and features of words/phrases for new findings listed by the Chinese scientific articles. [Limitations] We only examined papers from the field of natural sciences. More comparative research is needed to analyze papers from other areas. [Conclusions] Annotating corpus, counting frequency distribution statistics and analyzing of co-occurrence could effectively identify new findings from Chinese scientific articles.

[Objective] This paper presents a new model for public opinion monitoring system based on Hadoop to retrieve and analyze information from the micro-blog platforms. [Methods] We first surveyed the existing technology of the public opinion monitoring systems and proposed a new model with modified algorithm. Then, we built a big data analysis platform with Hadoop to examine the model’s feasibility through experimental simulations. [Results] The proposed model can detect and retrieve public opinion data effectively. [Limitations] The Hadoop cluster was relatively small. We did not compare our model with other clustering algorithms to discuss their advantages and disadvantages. [Conclusions] The proposed model can conduct public opinion analysis with micro-blog data and provide scientific information for the policy makers to improve crisis management.

[Objective] We explore the changing of consumer’s favorite brands by analyzing online product reviews from a popular E-commerce platform in China. [Methods] First, we built a fuzzy sentiment dictionary for online product reviews based on brand switching intention model. Second, we defined rules for a Fuzzy Inference System to calculate customer brand switching intention and switching types. [Results] We successfully extracted vague sentimental terms from the online product reviews, and then categorized consumers’ intentions. [Limitations] The fuzzy sentiment dictionary was built with complex rules and required many time consuming follow-up amendments. [Conclusions] The proposed model can provide decisive information for online marketing and early warning systems.

[Objective] To promote the knowledge reuse in product design and improve its efficiency. [Methods] First, we analyzed structure of the knowledge components in accordance with the knowledge reuse requirements of product design activities. Second, we proposed a framework for product design system based on knowledge components. Finally we implemented the reuse of knowledge components with the help of related computer programs. [Results] We constructed the knowledge components for barrel products, and then established a designing system based on the new knowledge components. [Limitations] The new system had certain limits, and could only be applied to the same or similar products. More research is needed to examine the universality of the knowledge components, the adaptability of the system, as well as the transferability of knowledge. [Conclusions] The new system based on knowledge components could effectively reuse the product design knowledge.

[Objective] This research aims to identify causal relationship of collaborative knowledge creation in the enterprise value chain. It also creates a model to improve the effectiveness of knowledge innovation. [Methods] We proposed a dynamics model for influencing factors of collaborative knowledge creation in enterprise value chain. Upon identifying these factors, we built the model based on findings from the causal relationship analysis. [Results] Using the Vensim PLE package, we constructed a simulation system with the proposed dynamics model, which further examined its effectiveness and sensitivity. This model represented the actual process of collaborative knowledge creation. [Limitations] We employed simulation data to examine the model. Further test is needed with actual data from the enterprise value chain. [Conclusions] We propose an optimization plan for collaborative knowledge creation based on the simulation results, aiming to improve its effectiveness in the enterprise value chain.

[Objective] This paper develops a parsing and indexing system for the WARC files, which fully exploits the value of Web archives from scientific institutions. [Context] The WARC files have been widely used in digital curation. However, the existing full-text indexing tools cannot satisfy the diversified needs of the WARC searchers. [Methods] We employed a modular scheme to parse the WARC files. Upon analyzing popular indexing tools, developed a new full-text indexing system based on the Solr platform. [Results] The new system effectively indexed the Web archives. Users could search information from different perspective, such as the subject category, resource type, and archived time, etc. [Conclusions] The new system indexes rich Web archives from international institutions, and improves the efficiency of users’ information retrieval activities.

[Objective] This paper designs and implements a personalized library usage data management system for the graduates based on their schools or departments. [Context] This new system helps the university launch new services for the graduating students. [Methods] We created a database to manage users’ data from different library departments. The public portal of this database was developed with JSP, HTML5, CSS and jQuery. [Results] The graduates could browse and print the records of their library visits, the circulation history, the seating records, booking details of the study rooms, etc. [Conclusions] This system shows the value of library data, and provides humanistic care to the graduates.