Entity Matching Search via Segmentation and Pruning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional approaches to matching query entities to target entities under value sum constraints become computationally intensive and impractical due to the exponential number of possible subsets, especially when the number of target entities exceeds 20, leading to excessive processing power and memory burdens.
Innovation Solution
A resource-efficient search method using decomposed searching and memory, where the search problem is broken down into smaller sub-problems, solved recursively, and solutions are memorized to avoid repetitive computations and early pruning of infeasible subsets, employing a hash table for memorization and a stack-based approach to efficiently explore the search space.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If exhaustive enumeration of all subsets is used to search for matching entities under value sum constraint, then complete solution is obtained, but computational complexity becomes exponential and processing resources are excessively consumed
Solution Approach 1:
The patent segments the exhaustive search space into multiple search branches by sorting target entities and systematically exploring subsets. Instead of enumerating all 2^N subsets, the method divides the search into manageable segments using a sorted order, allowing pruning of infeasible branches early in the process while maintaining completeness of the final solution.
Solution Approach 2:
The patent performs preliminary sorting of target entities by value before the search process. This preliminary action enables the subsequent search algorithm to efficiently prune infeasible subsets by comparing running sums against the target value, avoiding the need to explore all possible combinations and significantly reducing computational complexity.
2Reliability
If exhaustive enumeration of all subsets is performed, then all possible matches are identified, but memory consumption becomes prohibitive for large datasets
Solution Approach 1:
The patent segments the subset enumeration process into incremental steps, processing subsets in a systematic order rather than loading all 2^N subsets into memory simultaneously. By using a sorted approach with early pruning, the method maintains only the current search path in memory, reducing space complexity from exponential to polynomial.
Solution Approach 2:
The preliminary sorting of target entities enables memory-efficient search by allowing the algorithm to prune infeasible subsets early. This prevents the accumulation of infeasible subset data in memory, as branches that cannot lead to valid solutions are abandoned before consuming significant memory resources.
3Measurement precision
If traditional exhaustive search is used, then accurate matching under value sum constraint is achieved, but processing time becomes unacceptable for large numbers of target entities
Solution Approach 1:
The patent performs preliminary sorting of target entities by value in ascending order. This enables the search algorithm to systematically explore subsets and prune infeasible branches early by comparing running sums against the target value, ensuring that only subsets that could potentially satisfy the constraint are fully evaluated. This maintains accuracy while dramatically reducing processing time.
Solution Approach 2:
The search process is segmented into systematic exploration of subsets in sorted order, with early termination of infeasible branches. This segmentation allows the algorithm to maintain precision in identifying valid matches while avoiding the time-consuming enumeration of all possible subsets, achieving polynomial time complexity instead of exponential.
Data Source
AI summary
Methods, systems, and computer-readable storage media for receiving a set of inference results generated by a ML model, the inference results including a set of query entities and a set of target entities, each query entity having one or more target entities matched thereto by the ML model, processing the set of inference results to generate a set of matched sub-sets of target entities by executing a search over target entities in the set of target entities based on constraints, for each problem in a set of problems, providing the problem as a tuple including an index value representative of a target entity in the set of target entities and a value associated with the query entity, the value including a constraint relative to the query entity, and executing at least one task in response to one or more matched sub-sets in the set of matched sub-sets.