Knowledge discovery

What is it, how does it work, for whom is it of use, where can it be done?

Marko Grobelnik and Dunja Mladenic, J.Stefan Institute

Knowledge Discovery could be defined as a set of techniques coming mainly from the area of Artificial Intelligence but also borrowing important building blocks from other fields such as Statistics and Databases. The main goal of the whole area is to find useful pieces of knowledge within the data with none or little human involvement. Knowledge Discovery is useful for people interested in obtaining some useful information from large amount of data including data provided in different forms (databases, text, Internet pages, Web server log-files, pictures, video, sound) and data coming from different data sources. Some of the data formats are more appropriate for automatic data analysis and easier to handle than others. The usual data analysis methods assume that the data is stored in one or more tables, organized in a number of fields with a predefined range of possible values. Text Mining is a subfield of Knowledge Discovery offering methods capable of handling the text data in order to obtain some insights. One of the most popular applications of Text Mining is document categorization. Document categorization aims to classify documents into pre-defined categories or topic ontology based on their content. For instance, classifying Reuters news into one or more topic categories such as news talking about “acquisitions” or news talking about “finances” etc. Other important problems addressed in Text Mining include document clustering, visualization, search based on the content, automatic document summarization, construction of document topic ontology, document authorship detection, user profiling.

We can say that the users of Knowledge Discovery techniques include people from the institutions dealing with large databases or document management systems (eg., publishing houses, news industry, libraries, archives), team leaders, managers, government officials as well as practitioners. Knowledge Discovery is used to address problems from different areas including the areas where intensive research on developing and applying the Knowledge Discovery techniques is in the last years intensively supported by EU such as, Information gathering and filtering, Digital libraries, Semantic Web, Language and Multimedia Learning, Knowledge Management, E-commerceMarketing and user profiling, Business mapping, Competitor Analysis, Risk Management, E-science, E-learning, Computer Security and Data Privacy, Cultural Heritage, Health care.

CLOSE WINDOW TO EXIT