Co-click Query Similarity Scoring for Ad Targeting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current advertisement systems face challenges in identifying relevant keywords for search queries, as they often rely on exact matches or expanded matches that may not capture the full range of user intent, leading to suboptimal ad targeting.
Innovation Solution
The system generates feature vectors for queries based on related queries identified through co-click analysis, allowing for the determination of similarity between queries, which enables the association of one query as a keyword for another, thereby improving ad relevance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If exact match or expanded match is used to identify advertisements, then the identification process is simple, but the relevance of advertisements to user intent is insufficient
Solution Approach 1:
The system pre-computes feature vectors for queries by analyzing co-click data and generating related query information in advance. This preliminary processing enables the system to quickly determine ad relevance without performing complex real-time analysis, thus improving measurement precision while managing complexity
Solution Approach 2:
The patent introduces feature vectors as an intermediary representation between raw queries and advertisement matching. These feature vectors capture semantic relationships and user intent, serving as a bridge that improves ad relevance determination without requiring direct complex comparison of raw query terms
2Measurement precision
If co-click analysis and feature vectors are used to determine query similarity, then ad targeting precision is improved, but computational complexity increases
Solution Approach 1:
The system performs co-click analysis and feature vector generation as preliminary offline processes. By pre-computing these complex features before actual ad serving, the system achieves high measurement precision for query similarity while avoiding the burden of real-time computational complexity
Solution Approach 2:
The patent creates feature vector copies that represent queries in a simplified numerical format. Instead of performing complex semantic analysis during ad serving, the system uses these pre-generated vector copies for efficient similarity computation, reducing system complexity while maintaining precision
Data Source
AI summary
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for determining similarities between queries. A first query and a second query are identified, and first related queries associated with the first query and second related queries associated with the second query are also identified. One or more first features for the first query are identified, and one or more second features for the second query are also identified. A first score is calculated for each first feature based on the relatedness weight associated with the respective first related query including the term of the respective first feature, and a second score is calculated for each second feature based on the relatedness weight associated with the respective second related query including the term of the respective second feature. A similarity of the first query and the second query is determined based on the first score and the second score.