Document Partial Membership Classification for Search and Security
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
As users access and store increasing numbers of documents across various contexts, finding a specific document becomes challenging due to inadequate classification and security measures, leading to potential unauthorized access.
Innovation Solution
A system that classifies documents based on various characteristics and assigns them partial or full membership in multiple communities, using machine learning techniques and fuzzy logic to determine appropriate storage locations, ensuring compliance with corporate and regulatory guidelines.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If documents are stored in multiple locations without classification, then storage accessibility is improved, but document search relevance deteriorates
Solution Approach 1:
The patent segments documents into multiple communities based on their characteristics and assigns partial membership scores. Documents are divided into different communities (e.g., personal, work, sensitive) with varying degrees of membership, enabling organized storage across multiple locations while maintaining searchability through weighted associations.
Solution Approach 2:
The patent applies local quality by assigning different membership scores to different communities for each document. Each document has a unique profile of membership scores across multiple communities, allowing precise localization of documents to specific storage locations based on their characteristics and the user's needs.
2Device complexity
If documents are not classified by characteristics, then storage simplicity is improved, but security against unauthorized access deteriorates
Solution Approach 1:
The patent segments documents into different communities based on sensitivity and characteristics. Sensitive documents are assigned to specific communities with restricted access, while less sensitive documents remain in general accessible locations. This segmentation maintains security without requiring complete system complexity.
Solution Approach 2:
The patent changes the parameter of document classification by introducing membership scores that quantify the degree to which a document belongs to a community. This parameter enables automated security decisions based on numerical thresholds, balancing security requirements with system simplicity.
3Device complexity
If documents are assigned to single community only, then classification simplicity is improved, but adaptability to multiple contexts deteriorates
Solution Approach 1:
The patent applies partial action by assigning documents to multiple communities with partial membership scores rather than requiring full membership in a single community. This allows documents to be partially associated with multiple contexts (e.g., a document can be 70% work-related and 30% personal), enabling flexible multi-context retrieval without overwhelming complexity.
Data Source
AI summary
Example implementations relate to assigning a document to partial membership in communities. In example implementations, a detected feature of a document may be compared with a training pattern. Based on the comparison, the document may be assigned partial membership in a first community and partial membership in a second community.