| Abstract [eng] |
Current information systems are made to update their data automatically, without any help from humans, but the problem of incorrect data recognition in semi-structured web pages arises. In this thesis you can find an algorithm, which was made to monitor data that is available from a web page. This algorithm is almost non-dependant on web page structure and does data recognition based on the data that is already known. Using this algorithm model a tool was created to monitor currency exchange rates. This tool performs on the same level as other available tools in terms of precision and accuracy of collected data and it also reacts to changes of the data source structure. |
