Effect of temporal relationships in associative rule mining for web log data

Area of Science:

• Computer Science
• Data Mining
• Web Analytics

Background:

• Web applications generate vast amounts of diverse web log data.
• Traditional rule mining methods may not fully leverage temporal information in web logs.

Purpose of the Study:

• To investigate the impact of temporal attributes on relational rule mining for web log data.
• To compare temporal rule mining with classical approaches like Apriori and FP-Growth.

Main Methods:

• Incorporating temporal characteristics into the rule mining process.
• Analyzing the effects of various temporal parameters.
• Comparing rule sets generated by temporal relational rule mining against classical algorithms.

Main Results:

• Temporal relational rule mining generates a smaller set of rules compared to classical methods.
• The quality of rules generated by the temporal approach is comparable to traditional algorithms.
• Incorporating time as an attribute refines the rule discovery process.

Conclusions:

• Temporal attributes significantly influence relational rule mining on web log data.
• The temporal approach offers a more focused and efficient method for web log analysis.
• This method provides comparable rule quality with reduced output size.