Unsupervised User Segment Discovery via Session Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Discovering important patterns of user activity on websites is challenging due to unknown key patterns and difficulties in identifying, evaluating, and describing usage effectively for website stakeholders.
Innovation Solution
An unsupervised user segment discovery system that identifies important segments and patterns within user data without labels, establishes key performance indicators, and automatically generates descriptions for each segment, using session representation clustering and natural language descriptions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If traditional user activity recording techniques are used to generate event streams and analytics, then user behavior data can be collected, but discovering important patterns becomes very challenging
Solution Approach 1:
The system enables unsupervised discovery of user segments and patterns automatically from event streams without requiring manual labeling or supervision. The clustering algorithms self-organize the data to reveal hidden patterns, allowing the system to serve its own pattern discovery needs without external intervention.
Solution Approach 2:
The patent introduces session representations as an intermediary layer between raw event streams and pattern discovery. These representations transform complex user interaction sequences into structured formats that can be effectively clustered, serving as a mediator that simplifies the pattern discovery process.
2Adaptability or versatility
If user segments are discovered without labels or ground truths, then the system can find hidden patterns, but it becomes difficult to evaluate and describe the segments effectively
Solution Approach 1:
The system employs feedback mechanisms where cluster validation metrics and segment characteristics are continuously evaluated. The clustering process receives feedback about segment quality and adjusts parameters to improve evaluation accuracy, ensuring that discovered segments are meaningful and can be effectively described.
Solution Approach 2:
The patent utilizes parameter changes in clustering algorithms to optimize segment discovery. By adjusting clustering parameters and evaluating different configurations, the system can identify segments that best represent underlying user behavior patterns while maintaining evaluability and describability.
3Measurement precision
If complex clustering algorithms are used to identify user segments, then more accurate patterns can be discovered, but the system requires more computational resources and time
Solution Approach 1:
The patent segments the user base into distinct clusters based on behavior patterns, processing data in manageable groups rather than analyzing all user data simultaneously. This segmentation approach enables accurate pattern discovery while reducing computational complexity and improving processing efficiency.
Solution Approach 2:
The system performs preliminary actions by pre-processing event streams into session representations before clustering. This preliminary organization of data into meaningful structures reduces the computational burden during the actual clustering process, enabling faster and more efficient segment discovery without sacrificing accuracy.
Data Source
AI summary
A computing system generates, for each of a plurality of sessions, a session representation indicating the order through which a user navigated through a plurality of webpages during the session, and generates for each pair of session representations, a score indicating how similar the session representations in each pair are to each other. The computing system further generates a plurality of clusters by clustering the session representations based on the score for each pair of session representations, determines an optimal number of clusters based on key performance indicators corresponding to each cluster of the plurality of clusters, and generates a natural language description for each cluster of the optimal number of clusters based on the key performance indicators associated with each cluster.