1129 Analysis of Research Hotspots in the Field of Global Medical Research Based on Natural Index Hot! Yang Ying, Xu Dan, Chen Sisi, Han Shuang, Xu Shuang DOI: 10.3772/j.issn.1000-0135.2019.11.001 Since the Natural Index was proposed in 2014, research has progressively increased on existing indicators, but scientometric analysis based on basic data is relatively scarce. Scientific quantitative analysis is based on publications in Natural Index journals in the field of global medical research. Based on the theories of co-word analysis and co-occurrence network, scientific measurement tools such as BICOMB and Ucinet are used to analyze the social network structure of research hotspots of the past two years. Additionally, the gCluto software is used for biclustering the word-article matrix to explore research frontiers. This research reveals the hotspots of global medical research, provides a reference for the development direction of the field, and analyzes the advantages and disadvantages of the Natural Index in the evaluation of scientific research. 2019 Vol. 38 (11): 1129-1137 [Abstract] ( 283 ) HTML (95 KB)  PDF (7240 KB)  ( 415 )

1138 Research on the Influence of Social Interaction on User Perception and Information Adoption of the Recommendation System Hot! Li Zhi, Sun Rui DOI: 10.3772/j.issn.1000-0135.2019.11.002 The personalized recommendation system (PRS) generates recommendation information by considering the preferences of target users and similar users. In this study, based on the situational experiment analysis, participantsevaluation of PRS is obtained based on social interaction by changing the level of social interaction (social reference and self-reference) and using six PRSs for the application to establish the experimental operation under the Web. Moreover, the experimental data are analyzed and processed by SPSS23.0 and Smart PLS2.0 software. The results indicate that the social interaction environment significantly improved the perceived accuracy and novelty of PRS. The results also confirm the positive impact of perceived accuracy and novelty on user satisfaction and that of satisfaction and perceived novelty on information adoption. In addition, the research verifies the mediating effect of perception accuracy, novelty perception, and satisfaction. This study aims to explore the influence of social interaction on perceived accuracy and novelty, which, in turn, affect satisfaction and information adoption. By integrating the functions of PRS and social interaction, we can improve our understanding of social cognitive processes related to PRS user perceptions. 2019 Vol. 38 (11): 1138-1149 [Abstract] ( 276 ) HTML (155 KB)  PDF (1083 KB)  ( 771 )

1150 Research on Altmetrics Evaluation Model of Academic Conference Based on GBDT Hot! Zhang Yang, Ye Yue, Zhang Zongxiang, She Fang, Chen Xiyu DOI: 10.3772/j.issn.1000-0135.2019.11.003 In order to spread and share academic achievements in a certain field, academic conferences serve as an inevitable and important part in the development of that subject field. In certain cases, academic conferences are favored by researchers because of their timely dissemination of knowledge. This paper collects relevant data from altmetrics indicators—Altmetrics. com and PlumX—on the international conferences on artificial intelligence from 2007 to 2014, with the help of descriptive statistics and gradient lifting decision tree. Using indicators screening, data imbalance processing, and model optimization, a meeting evaluation model based on the gradient lifting decision tree was formed. This model combines popular machine learning models with latest altmetrics indexes. It can effectively cover the deficiencies of traditional informetrics indicators and improve the accuracy of the conference evaluation model, which can enrich related research on conference evaluation and become a reference for future research. 2019 Vol. 38 (11): 1150-1159 [Abstract] ( 245 ) HTML (115 KB)  PDF (1935 KB)  ( 633 )

1160 An Extended Belief Network Retrieval Model Based on Document Relationships Hot! Xu Jianmin, He Dandan, Wu Shufang DOI: 10.3772/j.issn.1000-0135.2019.11.004 The performance of a retrieval model can be improved by using relationships among documents reasonably. To solve the problem of a basic retrieval model that does not use document relationships, an extended belief network retrieval model with two layers of document nodes is proposed. The topology structure and probability inference of the extended model are given. In the topology structure, relationships between items like terms and queries, terms and documents, and two layers of documents, are indicated by arcs. The relationship between documents is determined by the similarity of the documents. In the probability inference, the retrieval probability is made more accurate by using document similarity and the number of parent documents to modify the original probability. In experiments, the value of discounted cumulative gain and the precision-recall curve are introduced to attest to the performance of our proposed extended model. The results show that the extended model makes the ranking of related documents more reasonable and improves the precision under the premise of guaranteeing recall. 2019 Vol. 38 (11): 1160-1165 [Abstract] ( 172 ) HTML (149 KB)  PDF (1064 KB)  ( 449 )

1166 Multi-source Information Fusion Analysis for Emerging Technology Development Trend Identification, Using Blockchain as an Example Hot! Zhang Weichong, Wang Fang, Zhao Hong DOI: 10.3772/j.issn.1000-0135.2019.11.005 The scientific literature is constantly being enriched, and has become valuable quantitative analysis data. The information fusion analysis of different sources and different types of scientific and technical literature can provide powerful information support for comprehensively revealing the development status and trends of emerging technologies. The efficient acquisition of topics from multi-source heterogeneous data is a technical difficulty in solving the problem of “subject” measurement entities in multi-source information fusion. This article is aimed at studying seven different scientific and technical literature types: patents, journal articles, dissertations, conference papers, books, funding projects, and industry reports. A summary-based topic analysis method is proposed. The topic words are obtained from multi-source heterogeneous texts, and data fusion and topic association analysis are performed. The results are effective and efficient, which provides a reference for solving the problem. In the experiment, blockchain is taken as an example. Based on data fusion, sequential association analysis and topic association analysis are carried out to reveal the development of blockchain technology. The results show that the method effectively reveals the production process, the theme diffusion, and the evolution trajectory of blockchain technology innovation in the scientific literature. 2019 Vol. 38 (11): 1166-1176 [Abstract] ( 295 ) HTML (95 KB)  PDF (3129 KB)  ( 1363 )

1177 Research on Core Word Extraction Algorithm Based on Contextual Concept Hot! Shi Jin, Han Jin, Zhao Xiaoke, Liu Qianli DOI: 10.3772/j.issn.1000-0135.2019.11.006 As there has been little domestic or international research on the context core word extraction algorithm—researchers have mainly focused on the keyword extraction algorithm—this paper proposes a context-based dependency grammar analysis algorithm. Right at the outset, this paper proves the equivalency of the dependency parsing problem to the splitting of a sentence to obtain the minimum scale context and to find the core words in the minimum scale context. In order to solve these problems, this paper proposes two context core word solvingthe context core word extraction algorithm based on entropy comparison; and the context core word extraction algorithm based on the sum of indegrees comparison, and a minimum context solving algorithm is proposed to construct a dependency grammar tree. Data from 1152 valid papers in the Journal of the China Society for Scientific and Technical Information from 2007 to 2018 was collected for testing, which was compared with the keywords extracted by the classic keyword extraction algorithms TF/IDF, Text-Rank, and LDA. The experimental results show that the context-based dependency grammar analysis algorithm has a positive effect on the extraction of keywords. 2019 Vol. 38 (11): 1177-1186 [Abstract] ( 240 ) HTML (150 KB)  PDF (1079 KB)  ( 626 )

1187 The “Expected-Actual-Extended Application” Chain Model, and an Empirical Study of Large Scientific Facilities, Taking SACLA as an Example Hot! Guo Shijie, Wang Xuezhao, Han Tao, Wei Ren, Dong Lu, Li Yizhan, Li Zexia DOI: 10.3772/j.issn.1000-0135.2019.11.007 To reveal the potential applications of large scientific facilities, topic analyses on three levels (conceived application, actual application, and extended application) are proposed, using three kinds of data (experimental research proposals, scientific results, and citations in papers), with the analysis process including natural language processing, co-occurrence clustering, and bibliographic coupling. On this basis, an “expected-actual-extended application” chain model of large scientific facilities is summarized. The Spring-8 Compact Free Electron Laser (SACLA), a Japanese hard X-ray free electron laser, is studied as an example. The potential applications of SACLA in various scientific and technical fields are then discussed. 2019 Vol. 38 (11): 1187-1199 [Abstract] ( 263 ) HTML (91 KB)  PDF (9096 KB)  ( 508 )

1200 Research on the Discovery of Entity Relationships in Subdivided Domains under the Guidance of a Small-scale Knowledge Base Hot! Chen Guo, Xu Tianxiang DOI: 10.3772/j.issn.1000-0135.2019.11.008 The acquisition of entity relationships in subdivided domains is a key issue for deepening and generalizing applications of knowledge engineering. In order to tackle the core problem of heavy reliance on manually annotated corpus at present, a natural solution is to use the existing (or low-cost) knowledge base in the subdivided domains as a guide. In contrast to the general knowledge base, the domain knowledge base is often small. This means it is necessary to not only use the ready-made knowledge content, but also to fully explore the “domain meta-knowledge” contained in the domain knowledge base. This paper proposes a subdivided domain entity relationship discovery scheme that combines domain meta-knowledge and a word embedding vector analogy. First, this paper describes the entity relationship constraints of a specific subdivided domain based on the existing knowledge base, such as the symptom representation relationship, which consists of <disease, symptom> entity pairs. Secondly, the word embedding vector of the domain entity is calculated according to the corresponding domain corpus. Following this, the positive and negative case vector benchmarks of various relational word embedded analogies are learned to provide a small number of high-quality entity relationships in the knowledge base, with the entity relationship classifier then trained based on this. Finally, for a given domain entity, by combining relational constraints, word embedding similarity, and word embedding analogy results, the entities that form different types of relationships are obtained. Taking the cardiovascular data as an example, a small amount of domain knowledge extracted from the encyclopedia can be used to obtain a better entity relationship recognition effect. 2019 Vol. 38 (11): 1200-1211 [Abstract] ( 212 ) HTML (132 KB)  PDF (2432 KB)  ( 589 )

1212 Rethinking the Foundational Principles of Information Science Hot! Yang Jianlin DOI: 10.3772/j.issn.1000-0135.2019.11.009 The application frequency of the foundational principles of information science is generally low. In this article, the foundational principles of existing information science are reviewed and analyzed; further, the rationality and deficiencies of certain foundational principles are examined. The article identifies some irrationality regarding the existing foundational principles of information science. Explanations of the principles are too simple. Some foundational principles directly regard some phenomena or processing steps in the intelligence process as principle. Some foundational principles lack supporting explanations entirely or have explanations that are not sufficiently thorough. For example, the logarithmic perspective principle lacks a sufficient scientific basis and practical case; it also lacks the foundational principles of intelligence in the core concept. This article includes four suggestions for reconstructing the foundational principles of information science. First, a set of foundational principles based on the intelligence process is proposed. Second, a set of foundational principles based on a group of intelligence behaviors in the intelligence process is also proposed. Third, valuable ideas from the relevant expositions of the paradigms in information science are considered. Finally, foundational principles derived from laws in the field of information science have been eliminated. 2019 Vol. 38 (11): 1212-1221 [Abstract] ( 305 ) HTML (78 KB)  PDF (695 KB)  ( 932 )

1222 A Review of Word Representation Learning Hot! Pan Jun, Wu Zongda DOI: 10.3772/j.issn.1000-0135.2019.11.010 Word representation that reflects semantic meaning is fundamental to natural language understanding tasks. The traditional method of encoding a word through a semantic dictionary is impractical due to the high construction cost, and one-hot representation suffers from various defects, such as high dimension and data sparsity. Distributed word representation，which projects the words into vectors in a low-dimensional real-valued space, can capture the semantic relatedness between the words and has been widely used in many NLP tasks. In this paper, we present an in-depth study of word representation learning methods from the perspectives of input data, learning objectives, and optimization algorithms, focusing on the theoretical basis, key techniques, evaluation methods, and application fields. We then summarize the main challenges and the latest advances in this research field, and we finally discuss possible future work in the field. 2019 Vol. 38 (11): 1222-1240 [Abstract] ( 205 ) HTML (402 KB)  PDF (2532 KB)  ( 1176 )