Automatic Identification of Time and Subject Related Patterns in Large Collection of Scientific Publications

In the past decades, we have witnessed a dynamic development of scientific literature. The most important changes of this phenomenon relate to: (1) continuous increase in the number of published papers; (2) rapid development of electronic means of publication; (3) increased access to publications.The above trends are to be assessed positively. We must not, however, forget that the ongoing changes bring about certain difficulties that are new to us. Some of these difficulties may concern our inability to get acquainted with the most up-to-date issues because it is impossible to analyze the flood of information represented in numerous publications. The way to solve this problem may be application of automatic tools for information retrieval from text documents and application of data analysis methods that would enable a precise classification of documents based on the facts contained in them.Several methods connected with time and subject-related classification of text documentation will be presented as well as an attempt of their evaluation. As a result, key subjects as well as the trends that shape the analyzed research discipline will be identified. Papers published in MIS Quarterly in the period starting from 1977 until present will be used as the empirical data set.