Composite Scoring for Electronic Discovery Document Retrieval
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current document retrieval methods in electronic discovery are inefficient as they deliver binary results from multiple search methods, requiring users to review numerous irrelevant documents, leading to increased time and costs due to the lack of a comprehensive scoring system and coordination of results across different methodologies.
Innovation Solution
A method that coordinates results across multiple document retrieval systems by calculating a composite score for each document based on rankings from various search methods, such as exact, fuzzy, stemming, concept, and cognitive searches, to prioritize documents for review, thereby reducing the number of irrelevant documents presented to the user.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple document retrieval methods are used to improve recall rate, then more relevant documents are identified, but the number of false positives increases requiring more human review
Solution Approach 1:
The patent combines multiple document retrieval methods (keyword searching, fuzzy searching, stemming searching, concept searching, and cognitive searching) into a unified system that produces a single ranked list of documents. Each method contributes to a composite relevance score, merging their results to achieve high recall while maintaining precision through coordinated ranking.
Solution Approach 2:
The system changes the parameter of document ranking from binary classification to continuous composite scoring. By calculating composite scores that reflect overall relevance across multiple search dimensions, the system can prioritize documents more effectively, presenting the most relevant ones first while still capturing all potentially relevant documents.
2Measurement precision
If comprehensive document review is performed to improve precision, then fewer false positives remain, but time and cost increase significantly
Solution Approach 1:
The system performs preliminary automated document retrieval and ranking before human review. By pre-calculating composite scores and ranking documents in advance, the system prepares a prioritized list that guides human reviewers to examine the most relevant documents first, reducing the overall time required for accurate review.
Solution Approach 2:
The composite scoring system acts as an intermediary between automated retrieval and human review. It translates multiple search results into a unified ranking that mediates the transition from raw data to human analysis, optimizing the workflow by presenting documents in order of predicted relevance.
3Ease of manufacture
If binary classification results are used from multiple search methods, then implementation is simple, but coordination of results across methodologies is lacking
Solution Approach 1:
The system creates a universal composite scoring framework that can accommodate multiple different search methodologies. Each search method contributes to the same unified scoring system, allowing the framework to handle diverse retrieval approaches while producing coordinated, comparable results across all methods.
Data Source
AI summary
A method that includes obtaining results from multiple operations (e.g., searches) for a plurality of items (e.g., documents), determining a composite score for each of the plurality of items based on those of the results obtained for the item, and displaying information related to the composite score obtained for at least a portion of the plurality of items in a graphical user interface.