A Resourceful Info-Gain System Designed for Acquiring Frequent Sequential Traversal Outlines from Web Logs based on Dynamic Weight Constraint

Numerous methods for mining frequent sequential traversal trends have been created. These algorithms mine the set of often occurring subsequences traversal trends in a session database that fulfill a minimal support criterion. Nevertheless, prior frequent sequential traversal pattern mining methods assign a uniform weight to sequential traversal trends, despite the fact that the pages inside these trends varied in significance and weight. Another major issue with the majority of frequently occurring sequential traversal pattern mining techniques is that, when a minimum support is lowered, they generate a lot of sequential traversal patterns and do not offer any other options for adjusting the number of sequential traversal patterns besides raising the minimum support. Our proposal for a frequent sequential traversal pattern mining method with dynamic weight constraint is presented in this work. Our major strategy is to preserve the downward closure feature while incorporating the weight limitations into the sequential traversal trend. To preserve the downward closure property, a weight range is established, pages are assigned varying weights, and traversal sequences designate a minimum and maximum weight. A highest and lowest weight in the session database are used to trim rare sequential traversal subsequences during database scanning so that the downward closure attribute may be preserved. By varying the weight range of pages and sequence, our approach generates a few but significant sequential traversal trends in session databases with a low minimum support. For sequential trends, the most often used metrics are confidence and backing. When sub-trends are provided, the confidence assesses trends frequencies while the support evaluates trends frequencies. For some applications, these factors have significance and importance. To assess the level of surprise of the trends, the information gain metric—which is frequently employed in the field of information theory— might be helpful. The goal is to identify a group of trends whose information gain exceeds the minimal criterion.