Content Selection Using Precomputed Term-Content Records
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In online advertising, existing methods struggle to effectively target and select relevant advertisements for publication on publisher pages, often resulting in inefficient distribution and mismatched content.
Innovation Solution
A computer-implemented method that uses a record to associate terms with corresponding advertisements in a repository, determining queries based on statistical correlations to select relevant content, considering restrictions such as geographical and language limitations, to improve the selection and distribution of advertisements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional keyword matching is used to select advertisements, then the selection process is simple, but the relevance and targeting accuracy of advertisements deteriorates
Solution Approach 1:
The system pre-computes term-content correspondence information and stores it in a record before actual advertisement selection. This preliminary action includes analyzing content items, identifying terms, and pre-establishing matching relationships, so that when advertisement selection is needed, the system can quickly retrieve pre-computed information rather than performing complex analysis in real-time.
Solution Approach 2:
The patent introduces an intermediary record (lookup table) that mediates between terms and content items. This record stores pre-computed correspondence information, acting as a buffer that simplifies the selection process while maintaining high relevance. The intermediary structure allows the system to balance accuracy and complexity by storing detailed matching information without requiring complex processing during selection.
2Reliability
If comprehensive content analysis is performed to improve advertisement matching, then advertisement relevance improves, but processing time and computational resources increase
Solution Approach 1:
The system performs comprehensive content analysis in advance, extracting terms and establishing correspondence relationships between terms and content items before they are needed for advertisement selection. This pre-processing ensures high matching accuracy is achieved without incurring processing delays during actual advertisement delivery, as the analysis work is completed beforehand and stored for quick retrieval.
3Quantity of substance
If more terms are used in query formulation to improve content selection coverage, then the number of matching contents increases, but query complexity and processing overhead increase
Solution Approach 1:
The system pre-identifies and stores multiple terms associated with each content item in the record, along with their correspondence information. When forming queries, the system can retrieve these pre-identified terms and their match counts without performing complex term generation or analysis, thus increasing the number of matching contents while keeping query formulation simple.
4Measurement precision
If strict restrictions (geographical, language, content-based) are applied to improve advertisement targeting accuracy, then relevance improves, but the number of available advertisements decreases
Solution Approach 1:
The system pre-processes content items and advertisements, applying restrictions and filtering criteria in advance to identify matching pairs that satisfy all constraints. The record stores correspondence information only for pairs that meet the restrictions, so during advertisement selection, the system can quickly retrieve pre-filtered matches without needing to re-evaluate constraints, thus maintaining high targeting accuracy while ensuring sufficient available advertisements.
Data Source
AI summary
Among other disclosed subject matter, a computer-implemented method relating to selecting content for publication includes receiving a term to be used in selecting content for publication. The method includes obtaining information from a record using the received term, the information reflecting a correspondence between contents in a repository and the received term. The method includes determining, using at least the obtained information, a query to be performed on the repository for selecting at least part of the content.