The timely detection and rapid spread of new ideas and technological innovations, which are all called scientific and technological intelligence, plays a crucial role for scientific research. According to the Theory of Scientific Communication, scientific and technical intelligence consists of two parts: the formal part (scientific and technical literature) and the informal part (online scientific communication under the network environment). For the formal part, the rapid development of information technology and the wide application of Internet bring the new modes for literature retrieval, such as Open Access and online retrieval. However, as it is shown by the research report, current literature retrieval systems cannot solve the problem of interdisciplinary retrieval, so that the references are still restrained within the original areas. As a result, the spread and adoption of creative ideas and technological improvements are delayed. For the informal part, Web2.0 makes online public communication and discussion among scientists possible. However, since the discussions are self-organized without any central control, the useful information distribute in the network disorderly. So it is hard for these online discussions to play a role of providing real guidance to practical scientific research Therefore, it is crucial for scientific research to improve the efficiency and accuracy of scientific literature retrieval, as well as to obtain useful scientific intelligence from online exchanges. Basing on the existing works in the area of information retrieval and social network, we try to answer these questions with three aspects： 1. We study on the acquisition of scientific information. We fetch the literature and online scientific discussions from the major scientific Medias using the specifically designed Web crawlers. The literature directory is automatically constructed basing on keywords extracted from the documents, which overcomes the traditional expert directory depending on experience and can constantly update and expand. 2. We do researches on the classification algorithms and retrieval techniques of scientific information basing on the analysis to scientific relationship network. We propose a semantic-based text classification method basing on a Bayes network classification and also a similarity calculation method basing on the author cooperation network and literature citation network to reflect the relevance of authors and papers. The roles ...
修改评论