Table Component Feature Computation for Search Ranking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Structured data search systems face challenges in effectively indexing and ranking tables due to the diversity of content generation on the Internet, leading to less relevant search results and difficulty in determining the order of relevant tables among numerous matches.
Innovation Solution
Computing features of table components such as rows, columns, and cells to rank them, which can be aggregated for ranking tables, using a computer architecture that includes modules for table extraction, understanding, annotation, feature computation, and index building.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional search engine indexing methods are used on structured data, then the system can handle general web content, but it fails to effectively index and rank tables due to diversity in content generation
Solution Approach 1:
The patent applies parameter changes by computing multiple features (e.g., table quality score, relevance score, completeness score) from the raw structured data to transform diverse table representations into standardized comparable parameters. This allows reliable ranking across different data sources while maintaining adaptability to various table formats.
Solution Approach 2:
The patent segments the table ranking problem into multiple independent feature computations (structural features, content features, source features) that can be processed separately and then aggregated. This segmentation enables the system to handle diverse structured data effectively while maintaining consistent ranking criteria.
2Quantity of substance
If multiple portions of structured data are returned as matches, then comprehensive search results are provided, but there is no way to determine the order in which to present them
Solution Approach 1:
The patent performs preliminary action by computing all necessary features and scores for each matching table before the user needs to view results. The tables are pre-ranked based on aggregated features, so when results are presented, they are already in the optimal order without requiring additional user effort or system computation during result display.
Solution Approach 2:
The patent implements feedback mechanisms where user interactions with ranked results (clicks, views) are used to refine and update the feature computations and ranking criteria, making the ordering process increasingly accurate and easier over time.
3Measurement precision
If comprehensive features are computed for all table components, then accurate ranking is achieved, but the computing complexity increases
Solution Approach 1:
The patent applies partial action by computing only the most relevant features for each table based on its specific characteristics, rather than computing all possible features uniformly. This selective feature computation maintains ranking accuracy while reducing overall system complexity and computation resources required.
Data Source
AI summary
The present invention extends to methods, systems, and computer program products for computing features of structured data. Aspects of the invention include computing features of table components (e.g., of rows, columns, cells, etc.). Computed features can be used for ranking the table components. When aggregated, features for different components of a table can be used for ranking the table (e.g., a web table).