Analytical Name Search Using Bigram Bitmap Signatures
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Analytical name search engines are slower due to the need to compare query names with every name in the data list, despite being more accurate, indicating a need to enhance performance without sacrificing accuracy.
Innovation Solution
Creating query and data list name bitmap signatures using bigrams, calculating a query name bitmap range, and identifying a subset of data list names within this range to reduce comparisons, while optimally arranging bit positions to eliminate non-matching data list names.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If analytical name search compares query name to every name in the data list, then search accuracy is maintained, but search performance deteriorates
Solution Approach 1:
The patent segments the name comparison process into multiple stages: first using bitmap signatures for rapid filtering, then applying analytical comparison only to candidate names that pass the bitmap filter. This segmentation allows the system to maintain accuracy through analytical comparison while improving performance by eliminating non-matching names early using the bitmap preprocessing step.
Solution Approach 2:
The patent performs preliminary actions by creating bitmap signatures for all names in the data list before the actual search query is processed. These pre-computed bitmaps enable rapid candidate elimination during search execution, reducing the number of names that require full analytical comparison and thus improving search performance without sacrificing accuracy.
2Productivity
If bitmap signatures are used to reduce comparisons, then search performance improves, but search accuracy may deteriorate
Solution Approach 1:
The patent introduces bitmap signatures as an intermediary mechanism between the query and the full analytical comparison process. The bitmap acts as a mediator that quickly identifies candidate names worth further investigation, ensuring that no potential matches are lost while eliminating clearly non-matching names. This intermediary step maintains accuracy by preserving all true candidates for analytical comparison.
Solution Approach 2:
The patent applies different comparison strategies to different subsets of names: bitmap-based filtering is applied to the entire data list for rapid candidate elimination, while full analytical comparison is applied selectively only to the subset of names that pass the bitmap filter. This local quality approach ensures high-speed processing for the majority of names while maintaining high accuracy for the candidate subset.
3Measurement precision
If all bit positions are compared between bitmap signatures, then comparison accuracy is maintained, but processing time increases
Solution Approach 1:
The patent extracts and compares only the relevant bit positions from the bitmap signatures rather than comparing all bit positions. By identifying and comparing only the critical bits that determine match eligibility, the system reduces processing time while maintaining comparison accuracy for the candidate names that reach this stage.
Data Source
AI summary
A system creates a query name bitmap signature comprising at least one bigram for a query name. The system calculates a query name bitmap range for the query name based on a match threshold and a number of bigrams in the query name bitmap signature. The system creates a data list name bitmap signature comprising at least one bigram, for each of a plurality of data list names, and identifies a subset of the plurality of data list names. The subset comprises data list name bitmap signatures that are within the query name bitmap range. The system reduces a number of comparisons by comparing only the query name and each of the subset of the plurality of data list names when performing the analytical name search. The system also reduces the number of comparisons by reducing the number of bit positions that are compared.