Metadata-Driven Search Result Culling in Electronic Discovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional electronic discovery tools rely heavily on keyword searches, resulting in vast and unwieldy search results that often include irrelevant and unimportant files, making it difficult to identify relevant documents for legal proceedings.
Innovation Solution
A computer-implemented method that involves identifying data sources, mining metadata to generate non-content-based attributes, creating an index to categorize files, receiving user input to specify categories and keywords, and searching within those categories to provide refined results that include only relevant files.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If keyword searches are used to identify relevant files, then the search can cover all files in the data source, but the results include vast amounts of irrelevant and unimportant files
Solution Approach 1:
The patent segments the search process into two distinct phases: first generating metadata derivatives from file contents, then using these derivatives to filter and organize files into groups before final keyword searching. This segmentation allows the system to reduce the search space from all files to only those likely to be relevant, resolving the contradiction between comprehensive coverage and result relevance.
Solution Approach 2:
The patent performs preliminary actions by generating metadata derivatives and creating grouped indexes before the actual keyword search. Files are pre-processed to extract meaningful attributes and organized into relevance groups, so that when keyword searches are executed, they operate on a pre-filtered subset rather than the entire data source, thereby reducing irrelevant results.
2Reliability
If all files are searched for keywords, then no relevant files are missed, but the search process becomes time-consuming and inefficient
Solution Approach 1:
The patent applies partial action by searching only within pre-identified relevant groups rather than all files. The metadata derivative analysis identifies a subset of files that are most likely to contain relevant information, allowing the keyword search to be performed partially on the data source while maintaining high confidence that relevant files are not missed.
Solution Approach 2:
By performing preliminary metadata extraction and grouping before keyword searching, the system prepares the search space in advance. This preliminary organization allows the actual keyword search to proceed much faster since it operates on a reduced, pre-filtered set of files rather than the complete data source.
3Adaptability or versatility
If metadata mining is performed on all files, then comprehensive categorization is achieved, but the preprocessing time and computational resources increase
Solution Approach 1:
The patent segments metadata processing by focusing on extracting specific derivative attributes (metadata derivatives) that are most useful for relevance determination, rather than analyzing all possible metadata fields. This selective segmentation reduces processing time while maintaining categorization accuracy.
Solution Approach 2:
The patent transforms metadata parameters by deriving new attributes (metadata derivatives) that capture essential relevance information. This parameter transformation converts raw metadata into meaningful categorization criteria, achieving accurate file grouping without requiring exhaustive analysis of all metadata fields.
Data Source
AI summary
The disclosed computer-implemented method for culling search results in electronic discovery may include (1) identifying a data source that is subject to electronic discovery in connection with a lawsuit, (2) mining metadata of a plurality of files within the data source to generate metadata derivatives that describe non-content-based attributes of the files, (3) creating an index that groups the files into a plurality of categories representative of the metadata derivatives, (4) receiving, from a user involved in the electronic discovery, user input that specifies a subset of the categories within the index and a keyword, (5) searching the subset of categories within the index for a subset of the files whose contents include the keyword, and (6) providing, to the user, a result of the search that identifies the subset of files whose contents include the keyword. Various other methods, systems, and computer-readable media are also disclosed.