Sponsored Content Scoring Using Unique Query Term Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems face challenges in designing a scoring process for sponsored content that prevents inappropriately high scores due to multiple occurrences of matching query terms across various sets, leading to suboptimal content selection.
Innovation Solution
A method is implemented to detect unique matching query terms across multiple sets, adjusting the score of sponsored content items based on the number of unique matches, applying penalties or increases to ensure better quality selections.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If multiple sets of query terms are used to select sponsored content, then the coverage and versatility of content matching is improved, but the scoring accuracy deteriorates due to duplicate matching terms contributing multiple times
Solution Approach 1:
The patent extracts duplicate query terms that appear across multiple sets and removes their redundant contributions to the score. By identifying and isolating the duplicate matching condition, the system prevents the same term from inflating the score multiple times, thereby maintaining scoring accuracy while preserving the benefits of multi-set query term matching.
Solution Approach 2:
The patent applies different quality standards to different parts of the matching process. Specifically, it treats first-time matches differently from duplicate matches across query term sets. The scoring mechanism assigns appropriate weight only to unique matches, creating a localized adjustment that preserves overall matching versatility while correcting the specific accuracy issue in scoring calculation.
2Ease of operation
If a simple matching operation is applied across multiple query term sets, then the processing speed and ease of operation are improved, but the reliability of score calculation deteriorates
Solution Approach 1:
The patent performs preliminary identification of duplicate query terms before final score calculation. By detecting duplicate matching conditions in advance and adjusting the scoring logic accordingly, the system ensures reliable score calculation without requiring complex post-processing or retraction of scores, thus maintaining processing simplicity while improving reliability.
Data Source
AI summary
Multiple sets of words and phrases can be used as query terms for accessing sponsored content. Sponsored content is associated with item terms. The query terms are matched with the item terms. The result of the matching operation is a numerical score for each item of sponsored content that can be used to select sponsored content. Conditions present across the set of matched terms are detected. Such conditions can include the occurrence of a matched query term in multiple sets of query terms. This condition can be addressed by determining a number of unique matching query terms across all sets of query terms. The score given to an item of sponsored content can be adjusted if such a condition occurs. For example, the score could be adjusted according to a function of the number of unique matching query terms across all sets of query terms. If the number of matching query terms across all sets of query terms is low, such as one or two, then the function can define a penalty to be applied to the score given to the item of sponsored content. If the number of matching query terms across all sets of query terms is higher, such as three or more, then the function can define an adjustment to be applied to the score given to the item of sponsored content.