Example-Based Item Classification Rule Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for automatic classification of items, such as emails, require users to manually create classification rules, which is a cumbersome and time-consuming process that can interfere with productivity and deter users from utilizing automated classification benefits.
Innovation Solution
A method that generates classification rules based on user-selected examples by analyzing shared text-based features across a subset of items, allowing for automatic classification of similar items and enabling users to refine rules through iterative adjustments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If users manually create classification rules using a structured rule-creating utility, then classification accuracy can be achieved, but the process becomes cumbersome and time-consuming
Solution Approach 1:
The system performs preliminary analysis on a small example subset of items to pre-compute shared features and generate candidate classification rules before full deployment. This preliminary action on representative samples enables the system to establish classification logic without manually processing the entire item set, significantly reducing the time required while maintaining accuracy.
Solution Approach 2:
The system enables self-service classification rule generation by automatically analyzing user-selected example items, identifying shared features, and generating candidate rules without requiring users to manually configure complex rule parameters. The system serves itself by performing the analytical work that would otherwise require significant user time and effort.
2Measurement precision
If users manually create classification rules, then specific and accurate classification can be achieved, but user productivity is interfered with and adoption is reduced
Solution Approach 1:
The system applies partial action by analyzing only a small example subset of items rather than requiring comprehensive manual rule definition. This partial analysis on representative samples is sufficient to generate accurate classification rules, achieving the desired specificity while requiring minimal user time investment and maintaining productivity.
Solution Approach 2:
The system performs self-service by automatically analyzing example items and generating classification rules without requiring users to manually define classification criteria. This eliminates the productivity interference associated with manual rule creation while maintaining classification accuracy through automated feature analysis.
3Productivity
If automated classification is implemented without manual rule creation, then productivity is improved, but the ability to generate specific and accurate rules may be compromised
Solution Approach 1:
The system performs preliminary analysis on user-selected example items to pre-compute shared features and generate candidate classification rules. This preliminary action on representative samples ensures that the automated process captures the specific characteristics needed for accurate classification while maintaining high productivity through automation.
Solution Approach 2:
The system implements feedback by presenting generated candidate classification rules to users for review and validation. This feedback loop ensures that automated rule generation maintains accuracy by allowing users to verify that the automatically generated rules correctly capture the intended classification logic before deployment.
Data Source
AI summary
Item classification rules are created based on examples selected by a user, such as by selecting a subset of emails, and the rule is used across a larger set of items to obtain automatic classification of similar items according to the rule. Based on an analysis, a candidate classification rule is generated identifying text-based features shared among the items of the subset. The user can review the candidate rule as well as a resultant subset of items generated by the rule, and either accept the candidate rule or make an adjustment to the examples and then perform one or more iterations of the analysis to refine the rule. Adjustments can be made by removing items incorrectly included in a resultant subset and/or adding items incorrectly excluded from a resultant subset, and using the adjusted subset in a next iteration.