Context-Based Search Result Filtering for Precise Document Review
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The exponential growth of electronically stored information has led to significant costs and inefficiencies in document review processes, particularly in identifying responsive, privileged, and confidential documents, with existing methods struggling to balance recall and precision, leading to increased time and risk of disclosure.
Innovation Solution
A computer-based method for filtering search results by identifying and grouping search term results, allowing users to evaluate representative fragments for relevance, and removing or retaining results based on context, thereby reducing false positives and improving precision.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional document review methods are used to identify responsive, privileged, and confidential documents, then comprehensive review can be achieved, but the process becomes extremely time-consuming and costly
Solution Approach 1:
The patent segments the document review process into multiple prioritization levels (first level, second level, third level) with different review criteria and depth. This allows reviewers to process documents efficiently by applying appropriate review intensity at each level, reducing overall time while maintaining comprehensive coverage through the hierarchical structure.
Solution Approach 2:
The patent employs preliminary actions through automated search term generation and document prioritization before human review begins. The system pre-processes documents by identifying potential responsive, privileged, and confidential documents using search terms derived from case information, creating a prioritized list that reduces the time reviewers need to spend on manual assessment.
2Productivity
If search terms are used to identify responsive documents, then review efficiency can be improved, but precision and recall balance becomes difficult to achieve
Solution Approach 1:
The patent implements dynamic search term generation that adapts to the specific case information and document corpus. The system generates search terms based on the actual case facts, parties involved, and document characteristics, allowing the search strategy to evolve and optimize precision for each unique review scenario rather than using static search terms.
Solution Approach 2:
The patent incorporates feedback mechanisms where reviewers can indicate whether documents are responsive, privileged, or confidential, and this feedback is used to refine future search term generation. The system learns from reviewer decisions to improve the accuracy of document identification in subsequent review cycles, balancing precision and recall through iterative optimization.
3Reliability
If multiple review levels are conducted to ensure accuracy, then document classification reliability improves, but the process becomes more complex and time-consuming
Solution Approach 1:
The patent divides the complex review process into three distinct segmentation levels: first level for responsive document identification, second level for privilege and confidentiality assessment, and third level for final verification. Each level has specific review criteria and can be independently configured, making the overall complex process more manageable and less time-consuming through structured organization.
Solution Approach 2:
The patent applies local quality by assigning different review depths and criteria to different document types and categories at each level. Documents are evaluated with appropriate intensity based on their characteristics (e.g., privileged documents receive enhanced review at specific levels), optimizing accuracy for each local context while reducing unnecessary review effort for lower-risk documents.
Data Source
AI summary
A computer-based method is provided for filtering search results, including first identifying a plurality of documents for review and then identifying an initial plurality of search term results potentially relevant to a search query, each search term result of the initial plurality containing an occurrence of a primary search term of a plurality of primary search terms. The method then proceeds to group identical search term results to define a plurality of groups of search term results. Representative search term result of each group are presented for evaluation. Upon receiving an indication from a user that the representative search term result is to be removed, the method then removes all search term results of the corresponding group to define a modified plurality of search term results. Each document from which any search result of the modified plurality of search term results is drawn is defined as a potentially relevant document.