Rule Discovery System for Efficient CFD Enumeration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing CFD discovery algorithms struggle to efficiently enumerate rules that 'substantially hold' in databases, as they either only identify rules with full confidence or take excessively long to compute for large-scale databases due to combinatorial explosions.
Innovation Solution
A system and method that generate free itemsets with frequencies above a threshold, create rule candidates with confidence above a set value, and check for minimality to efficiently output minimal and non-redundant approximate constant CFDs, sorting and extracting them in ascending order of size to reduce computational time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing CFD discovery algorithms are used to enumerate rules that substantially hold in databases, then complete rule coverage can be achieved, but computational time becomes excessively long due to combinatorial explosions in large-scale databases
Solution Approach 1:
The patent segments the rule discovery process into distinct phases: generating free itemsets with frequencies above a threshold, creating rule candidates with confidence above a set value, and checking for minimality. This segmentation avoids exhaustive enumeration of all possible rules while still capturing the most significant ones, thereby reducing computational time without completely sacrificing rule coverage.
Solution Approach 2:
The patent introduces parameter thresholds (frequency threshold for free itemsets, confidence threshold for rule candidates) to filter the search space. By changing these parameters, the system can efficiently identify approximate constant CFDs that substantially hold in databases without examining all possible rules, thus resolving the contradiction between complete coverage and computational efficiency.
2Loss of information
If all CFD rule candidates are generated and checked for validity, then all valid rules can be identified, but the computational complexity increases significantly for large-scale databases
Solution Approach 1:
The patent performs preliminary actions by first generating free itemsets with frequencies above a threshold before creating rule candidates. This preliminary filtering reduces the number of candidates that need to be fully evaluated, thereby reducing computational complexity while still identifying all valid rules that meet the significance criteria.
Solution Approach 2:
Instead of generating all possible CFD rule candidates, the patent generates only those candidates that meet minimum frequency and confidence thresholds. This partial action approach focuses computational resources on the most promising candidates, reducing overall complexity while maintaining identification of valid rules.
3Loss of information
If minimal rules are identified to avoid redundancy, then only useful rules are output, but additional processing steps are required to check and enforce minimality
Solution Approach 1:
The patent extracts minimal rules from the set of generated rule candidates by checking and enforcing minimality conditions. This extraction process removes redundant rules that are subsumed by more general rules, ensuring that only the most useful and non-redundant rules are output, while the minimality check is integrated into the generation process to avoid excessive additional processing.
Data Source
AI summary
A system includes a free itemset generation unit to generate a set of free itemsets, each having a frequency in the database greater than or equal to a predetermined threshold value set in advance, a valid rule candidate generation unit to generate rule candidates and store the generated rule candidates, and a rule minimality decision unit to check minimality of each of the generated rule candidates and output the generated rule candidate to an output apparatus when the generated rule candidate is determined to be minimal.