Legacy Expert Rule Refinement Using Training Data Patterns
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Legacy expert systems struggle to effectively leverage the rapid increase in data and are not easily updated with new knowledge, data, and rules, leading to inefficiencies and a need for costly migrations to modern systems.
Innovation Solution
A system and method that utilizes relevant big data to evaluate and refine existing rules in legacy expert systems by identifying low-quality rules, applying machine learning to generate new rules and frequent sets, and integrating these into the system to enhance its performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If legacy expert systems use manually populated knowledge bases with rules from human experts, then the system maintains interpretability and structured reasoning, but the system cannot effectively leverage rapid increases in data and is difficult to update with new knowledge
Solution Approach 1:
The patent introduces an intermediary processing layer that sits between the training data and the expert system rules. This layer automatically generates candidate rules from training data, which are then evaluated against existing rules and quality metrics before being integrated. This intermediary mechanism enables the system to adapt to new data without directly modifying the core knowledge base structure, thus maintaining reliability while improving adaptability.
Solution Approach 2:
The system performs preliminary actions by pre-processing training data to generate candidate rules before they are evaluated and integrated into the expert system. This preliminary rule generation step allows the system to prepare for upcoming data updates and ensures that only high-quality rules that meet established criteria are added, maintaining rule quality while enabling continuous adaptation.
2Productivity
If legacy expert systems manually update knowledge bases with new rules, then the system maintains control over rule quality, but the process is time-consuming and requires significant expert involvement
Solution Approach 1:
The system implements self-service by automatically generating candidate rules from training data using machine learning algorithms, evaluating their quality against predefined metrics, and integrating approved rules without requiring manual expert intervention for each rule. This automated self-service mechanism dramatically increases rule update speed while maintaining quality control through systematic evaluation criteria.
Solution Approach 2:
The patent replaces the manual mechanical process of expert rule creation with an automated computational system. Machine learning algorithms substitute for human experts in generating and evaluating rules, transforming a time-consuming manual process into an efficient automated workflow that maintains quality through algorithmic evaluation rather than human judgment.
3Adaptability or versatility
If legacy expert systems migrate to modern machine learning systems, then the system can leverage large data volumes and update easily, but the cost of migration is significant and loss of interpretability occurs
Solution Approach 1:
The patent merges the advantages of both legacy and modern systems by combining automated rule generation from training data (modern ML capability) with the traditional expert system rule structure and evaluation framework. This hybrid approach allows the system to leverage large data volumes for rule generation while maintaining the structured, interpretable rule format of legacy systems, avoiding the need for complete migration to complex modern architectures.
4Measurement precision
If the system evaluates all existing rules against quality metrics, then the system identifies low-quality rules for refinement, but the computational resources required increase significantly
Solution Approach 1:
The system applies local quality by evaluating rule quality metrics selectively rather than uniformly across all rules. The evaluation process focuses computational resources on rules that are most likely to benefit from refinement or that show degradation in performance, rather than re-evaluating every rule with equal intensity. This localized approach maintains measurement precision for critical rules while reducing overall computational resource usage.
Data Source
AI summary
A system, computer program product, and method are presented for enriching existing legacy expert systems through refinement of existing rules therein. The method includes identifying a legacy expert system to be enriched, relevant training data, existing rules embedded within the legacy expert system, and, for each existing rule, one or more antecedent factors. The method also includes determining the existing rules do not meet a threshold value for established quality requirements, thereby identifying one or more low-quality rules. The method further includes identifying frequent sets of antecedent factors associated with each low-quality rule, where each frequent set of antecedent factors is established as a frequent set through at least meeting a threshold frequency of occurrence within the training data. The method also includes comparing the antecedent factors of each existing rule with the frequent sets of antecedent factors, and enriching the legacy expert system through refining the existing rules.