Hierarchical Website Mining for User Behavior Context
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional web usage analysis methods focus on page views without providing context on user navigation, making it difficult to capture and analyze user behavior patterns across multiple sessions, especially due to the large number of distinct user sessions.
Innovation Solution
A method that maps user sessions to a hierarchical website structure, determines similarity between sessions using metrics like edit distance, and clusters similar sessions to visualize trends and anomalies, while also segmenting users based on interest vectors and determining group closeness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If conventional web usage analysis methods are used to track page views, then the number of page views can be counted, but the context of user navigation and user behavior patterns cannot be captured
Solution Approach 1:
The patent implements a hierarchical website structure where webpages are nested within categories and subcategories. User sessions are mapped to this hierarchy, creating nested session representations that capture both specific page views and broader navigation patterns. This nested structure allows the system to retain detailed user behavior context while organizing data in a manageable hierarchical format.
Solution Approach 2:
The patent transforms flat page view data into multi-dimensional hierarchical session data by mapping user sessions to hierarchical website structures. This dimensionality change adds contextual layers (categories, subcategories) to the analysis, enabling capture of navigation patterns and user behavior context that simple page view counts cannot provide.
2Loss of information
If individual user sessions are tracked in detail, then user navigation patterns can be captured, but the large number of distinct sessions makes analysis and summarization difficult
Solution Approach 1:
The patent merges individual user sessions by mapping them to a common hierarchical website structure. Similar sessions are clustered together based on their hierarchical patterns, combining information from multiple sessions to identify common user behavior patterns. This merging process reduces the complexity of analyzing thousands of individual sessions while preserving navigation pattern information.
Solution Approach 2:
The patent transforms raw session data into hierarchical session representations by applying mapping parameters that convert flat webpage sequences into structured hierarchical paths. This parameter change organizes the data into a standardized format that facilitates efficient clustering and pattern recognition across large numbers of sessions.
3Loss of information
If hierarchical website structure mapping is implemented to capture user behavior patterns, then user navigation context is preserved, but the complexity of mapping and clustering sessions increases
Solution Approach 1:
The patent segments the website into a hierarchical structure with categories, subcategories, and individual pages. User sessions are segmented into hierarchical paths that reflect this structure. This segmentation breaks down the complex mapping task into manageable hierarchical levels, making the overall system more organized and easier to implement despite the added structural complexity.
Data Source
AI summary
Mining of websites that in one embodiment includes obtaining web usage data of user sessions of a website, wherein the website has a hierarchical structure with granular levels and has mapping from each webpage of the website into the hierarchical structure, mapping the user sessions to the hierarchical structure of the website resulting in hierarchical user sessions, initiating an edit distance metrics to determine similarity in the hierarchical user sessions, and clustering similar hierarchical user sessions into groups.