Методы и средства извлечения данных о персоналиях из авторефератов диссертаций

Abstract

The problem of extraction of data about a person from scarce data collection is studied. The data collections are public resources on the internet. When these data are collected and parsed they present additional value for users. Collecting such data is problematic because of it’s weak structure restrictions. Thus the system is suggested to automate information gathering and parsing. The initial task is to process personal data from thesis documents publicly available on the internet. This data presents information about scientists which can’t be obtained from other sources. The goal is to be able to make requests to the data having its semantics in mind and not only plain text. The prototype system is developed with PHP and XPath able to collect raw documents from digital repository of National Library of Ukraine by V. I. Vernadskiy. The system also extracts data from the collected documents and stores them locally in RDF data model suitable for specific data and for future exposition to the Semantic Web. The collection of more than 63000 documents was processed to test the system.