Figure 1. EduMON system’s overview Towards Automated Education Demand-Offer Information Monitoring: the Information Extraction Peteris Rudzajs Department of Systems Theory and Design Riga Technical University Riga, LATVIA Peteris.Rudzajs@rtu.lv Abstract— Dynamically changing work environment in knowledge economy causes the changes in knowledge requirements for labor. Therefore it becomes more and more important to be constantly aware of what education is currently demanded and what education is currently offered. The IT solution is vital to process various information sources, extract education information, and provide analysis mechanisms in automated manner. The education information extraction is detailed in this paper in the context of Education demand and offer information monitoring system by providing the workflow for semi-automatic skills extraction from the university course descriptions using developed term suggestion method. Keywords- education information; monitoring system; information extraction I. INTRODUCTION Rapid economic changes in the knowledge requirements for labor cause a necessity to monitor education demand and offer. Education demand and offer (d/o) can be described in terms of knowledge, skills, or competences required in work environment or obtained in university. For simplicity in this paper the terms "education information" and "skills" are used interchangeably to denote knowledge, skills, or competences. Education d/o monitoring both for university and industry can provide an insight in knowledge, skills, or competences currently demanded/offered in educational and industrial environment. To facilitate the monitoring of education d/o, university and industry should be provided with IT solutions that reduce the large amount of manual work currently necessary to overview, extract, and analyze the information from various education information sources (study and certification course descriptions, job advertisements etc.). In this case, the monitoring system design principles are applicable, as basic functionality of the monitoring systems includes gathering of source information, information processing and analysis to provide decision support information to users of the system. In the previous study [1] the architecture of the Education d/o information monitoring (EduMON) system had been proposed. The architecture (see Fig. 1) consists of various classes of services and supports the process from retrieving and extracting information from relevant sources to finding correspondence between education information in various sources. Education information when reflected in the information sources is represented using domain terminology. Thus the terms of the information source document related to domain terminology should be identified in the process of the information extraction. In the current study an example of semi-automated information extraction method is presented (method contributes to the development of Information extraction services of EduMON system). Solution is illustrated by terminology extraction from the university course descriptions based on the n-grams of text document, Dice's similarity metric and available Free On-Line Dictionary of Computing (FOLDOC) [2] consisting of more than 14 600 IT- related terms. Method is also applicable to extract terms from job advertisements and other relevant sources reflecting the education information in the area of Information and Communication Technology. II. EDUCATION INFORMATION EXTRACTION The extraction of terms related to domain terminology is necessary to provide the standardized reference of terms represented in text. Relevant terms extracted from various information sources can provide basis for applying the analysis mechanisms provided by the EduMON system's services of Analysis service class. Examples of analysis services can include the analysis of skills (represented as terms in information sources) demanded in job advertisements and offered in University courses. The process of term extraction can be either manual or semi-automatic. In manual extraction, the user manually annotates the document (e.g., University course or job advertisement) by linking the terms presented in document (further in the text - source terms) to the terms available in reference dictionary (further in the text - target terms). As the manual extraction of skills is time-consuming, the solution for semi-automatic extraction is proposed. Basic workflow for the