Citation Analysis And Topic Analysis Of Conference Papers In DBLP

Computer Science is a new-rising subject. It systematically implements information processing and computation application based on information and computation theories. Through decades of speedy development, lots of academic branches have emerged in computer science, clarifying as theoretical orientation and experimental orientation. Theoretical computer science includes theory of computation, information and coding theory, algorithms and data structures, programming language theory, formal methods. Applied computer science includes artificial intelligence, computer graphics and visualization, computer security and cryptography, computational science, computer networks, concurrent, parallel and distributed systems, databases, health informatics, information science and software engineering. Computer science has been an indispensable technological support for other subjects. Therefore, studies of current academic status and topics in computer science have becoming more and more important.DBLP(Digital Bibliography　& Library Project) is an online platform for papers searching in computer science field, supported by Universitat Trier, in Germany. It contains magnanimous journal papers and conference papers. We choose conferences papers in DBLP as research object, by conducting citation analysis and topic analysis, to get a summary about computer science development, academic status and topic trends.We conduct citation analysis from conference aspect and author aspect. According to different names and classifications of conferences, we choose paper numbers, citation numbers, citation number per paper, H-index as measurements. For authors, paper numbers, citation numbers, citation numbers per paper, H-index and G-index are universal indexs. We hope to understand the research direction of computer science and the distributions of most influential conferences, research institutions.Topic analysis concludes two parts, word frequency analysis and topic model analysis. Considering the mass data in DBLP conference papers, we choose abstracts of papers to conduct further analysis. Word frequency analysis is meant to do word frequency statistics of annual abstract sets. Sorting from high to low, we will acquire topics of that year. Topic model analysis based on LDA using latent semantic method to analysis abstracts. Unlike word frequency analysis,this method will produce one or more topics for a single paper.After citation analysis and topic analysis, computer science has two mains topic areas. One is the basic theory and application research, including programming language and software engineering, operating system, algorithm and theory. Global researchers have implemented lots of work to make improvements. Frontier researches, such as artificial intelligence, machine learning and pattern recognition and data mining, have been widely used in practice. Development of network has driving network research, information security and information retrieval to new height.