Representative Keyword Selection for Contextual Ad Serving
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current search engines face challenges in selecting contextually relevant advertisements for web pages, as they rely on manual keyword selection and lack efficient methods to determine the most representative keywords based on web page content.
Innovation Solution
A system and method for selecting representative keywords from an initial web page by analyzing the frequency of appearance and proper nouns, using these keywords to query a network, and ranking them for relevance, thereby selecting contextually relevant advertisements to serve alongside the web page.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual keyword selection is used for advertisement targeting, then advertisers can provide custom advertisements with associated keywords, but the system lacks efficient methods to determine the most representative keywords based on web page content
Solution Approach 1:
The patent segments the web page content analysis into multiple independent components: extracting candidate keywords from the web page, querying search engines for each candidate, analyzing query results, and ranking keywords based on multiple criteria. This segmentation allows each component to be processed independently and efficiently, resolving the contradiction between measurement precision and device complexity.
Solution Approach 2:
The patent performs preliminary actions by first extracting candidate keywords from the web page before conducting search engine queries. It also pre-ranks keywords based on frequency and other criteria before final selection. This preliminary processing reduces the complexity of the main keyword selection task while maintaining high accuracy.
2Measurement precision
If multiple keyword candidates are analyzed through search engine queries, then the most representative keywords can be identified, but the processing time and computational resources increase
Solution Approach 1:
The patent applies partial action by analyzing only a selected number of top-ranked keyword candidates rather than all possible keywords. It performs search engine queries on a limited set of candidates that meet certain frequency and relevance thresholds, thereby reducing processing time while maintaining keyword representativeness.
Solution Approach 2:
The patent changes parameters by adjusting the number of query results analyzed, the frequency thresholds for candidate selection, and the ranking weights. These parameter adjustments allow the system to balance between processing time and keyword accuracy based on specific requirements.
3Adaptability or versatility
If keyword frequency and proper nouns are used as selection criteria, then contextually relevant advertisements can be selected, but the system requires multiple querying and analyzing operations
Solution Approach 1:
The patent merges multiple selection criteria (keyword frequency, proper noun detection, query result analysis) into a unified ranking system. By combining these different approaches into a single comprehensive evaluation framework, the system achieves high adaptability for contextual relevance while improving overall productivity through integrated processing.
Solution Approach 2:
The patent creates a universal keyword selection system that can handle various types of web pages and advertisement types using the same core methodology. The multi-functional approach allows the system to adapt to different contexts while maintaining efficient processing through a standardized workflow.
Data Source
AI summary
Systems and methods in clued those for serving one or more advertisements contextually-relevant to an initial web page. Terms are selected from the initial web page that may be used as representative keywords to select advertisements that are contextually relevant to the initial web page. Keyword candidates are filtered via a series of operations that include: querying a network for web pages where the one or more keyword candidates are found, filtering the web pages based on time/date characteristics and a pre-defined number of web pages, analyzing the web pages using the one or more keyword candidates, and selecting certain keyword candidates to be representative keywords based on the analysis. The one or more representative keywords may be used to select one or more advertisements that may then be served with the initial web page.