Keyword Cluster Relevancy Scoring for Event-Based Content Serving
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems fail to effectively serve relevant third-party content to devices based on calendar events, as they rely solely on browsing history and lack the ability to adapt content relevance over time periods relative to event schedules.
Innovation Solution
A method and system that create a data structure with relevancy scores for keyword clusters related to event categories across various time periods, allowing for the selection of content based on the timing of events and the historical performance of content items, thereby adjusting content relevance dynamically.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If content serving systems rely solely on browsing history, then system complexity is reduced, but content relevancy to user needs deteriorates
Solution Approach 1:
The system performs preliminary actions by creating a data structure with pre-calculated relevancy scores for keyword clusters related to event categories across various time periods before content serving requests arrive. This allows the system to have relevant content ready in advance based on scheduled events, improving content relevancy without increasing real-time system complexity
Solution Approach 2:
The system implements dynamics by making content relevancy adaptive and changeable over time. Relevancy scores are generated for different time periods relative to events (e.g., before, during, after), allowing the system to dynamically adjust which keyword clusters are most relevant at any given moment, thereby improving content relevancy while maintaining manageable system complexity through structured temporal segmentation
2Loss of information
If the system generates relevancy scores for multiple time periods, then content relevancy improves, but computational resources increase
Solution Approach 1:
The system applies segmentation by dividing the time dimension into discrete time periods relative to events (e.g., pre-event, during-event, post-event periods). This segmentation allows computational resources to be distributed across different time segments rather than requiring all computations simultaneously, improving content relevancy while managing computational resource usage through structured temporal division
Solution Approach 2:
The system performs preliminary computation of relevancy scores for multiple time periods in advance, storing them in a data structure before content serving requests arrive. This preliminary action shifts computational load to off-peak times when events are scheduled, reducing real-time computational resource requirements while maintaining high content relevancy when needed
3Measurement precision
If the system uses historical content performance data, then content selection accuracy improves, but data processing complexity increases
Solution Approach 1:
The system performs preliminary analysis of historical content performance data to generate relevancy scores that are stored in a structured data structure. By pre-processing historical data and capturing essential performance patterns in advance, the system improves content selection accuracy without requiring complex real-time data processing during content serving operations
Solution Approach 2:
The system creates a simplified representation (copy) of historical content performance data in the form of pre-calculated relevancy scores for keyword clusters. This copying approach captures the essential information from complex historical data without requiring the system to re-process the full historical dataset during content serving, thereby improving selection accuracy while reducing data processing complexity
Data Source
AI summary
Systems and methods for identifying relevancy scores of a keyword cluster related to an event category for a given time period relative to an event of the event category. Processors identify an event and determine event parameters of the event. Processors determine keyword clusters associated with the event parameters for serving content. Processors generate a relevancy score for each of the keyword clusters over plurality of time periods relative to the event. The relevancy score of a first subset of the keyword clusters is higher during a first time period than a second time period and the relevancy score of a second subset of the keyword clusters is higher during the second time period than the first time period. Processors create, for the event, a data structure including the keyword clusters and the generated relevancy scores for each of the keyword clusters over the time periods relative to the event.