Abstract:
The World Wide Web is a popular and interactive tool for spreading information
today that is expanding rapidly. Web mining is the process of extracting patterns and
implicit data from artifacts or activities associated to the World Wide Web in order to
find relevant and potentially helpful patterns. This extracted data can also be utilized to
enhance web personalization, fraud detection, future prediction accessed by the user,
user profiling, and understanding of user web activity. Upcoming page prediction is
one example of how this information can be used to improve web usage mining. The
proposed system consists of three phases which are data pre-processing, pattern
discovery, and pattern analysis. Raw web log data may contain noise and impurities.
By using some data preprocessing techniques that noise will be removed. Data
preprocessing phase is the most important one because it makes the data with good
quality. In Pattern discovery phase, the users’ navigational pattern and rules are
extracted by using association rule algorithm. In this thesis, first preprocessing can be
done with data cleaning, user identification, page identification and session
identification. The objective of this thesis is to identify the frequent pattern from Online
Judge web log data of web server of the University by using the ECLAT algorithm. We
modified a very efficient ECLAT algorithm for matching interesting new patterns and
even used support and confidence to calculate interesting pattern measures.
