Zita Bosnjak, Olivera Grlјevic and Sasa Bosnjak

Faculty of Economics in Subotica, University of Novi Sad, Subotica, The Republic of Serbia

Much of one’s online behavior, including browsing, shopping, posting, is recorded in databases on companies’ computers on a daily basis. Those data sets are referred to as web data. The patterns which are the indicators of one’s interests, habits, preferences or behaviors are stored within those data. More useful than an individual indicator is when a company records data on all its users and when it gains an insight into their habits and tendencies. Detecting and interpreting such patterns can help managers to make informed decisions and serve their customers better. Utilizing data mining with respect to web data is said to turn them into web knowledge. The research study conducted in this paper demonstrates how data mining methods and models can be applied to the web-based forms of data, on the one hand, and what the implications of uncovering patterns in web content, the structure and their usage are for management.

Keywords: unstructured data, web mining, customer segmentation, on-line behavior modelling, collaborative filtering

JEL Classification: O33, L190, D85, Z13

Economic Horizons2019, 21(2), 173-189. Published online 23 August 2019