Geocoding Address Decomposition and Normalization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing geocoding processes face challenges in accurately matching user-input location information in varying formats with reference map data due to format inconsistencies and differences across countries, leading to inconsistent and unsuccessful geocoding results.
Innovation Solution
A geocoding method involving pre-decomposition, purification, decomposition, composition, and normalization stages to identify and rank address components, convert user-input formats, and match them with reference map data, using techniques such as removing unsupported characters, separating address components, and normalizing formats like highway names and abbreviations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional geocoding methods are used to match user input addresses with reference map data, then the process is simple and fast, but the accuracy deteriorates due to format inconsistencies and variations across countries
Solution Approach 1:
The address string is divided into multiple components (street number, street name, directionals, postal code, city, state, zip) and each component is processed independently through decomposition, purification, and composition stages. This segmentation allows the system to handle format variations in each component separately, improving overall matching accuracy without requiring complete reformulation of the entire address.
Solution Approach 2:
The system performs pre-decomposition and purification of the input address string before the actual matching process. Unsupported characters are removed, and the address is pre-segmented into components. This preliminary action prepares the data in advance, reducing the complexity of the main matching process and improving accuracy by eliminating formatting issues before they affect the match.
2Adaptability or versatility
If multiple address format variations are supported to improve adaptability, then the system becomes more versatile, but the matching process becomes more complex and less reliable
Solution Approach 1:
The system changes the parameters of the address string through normalization processes. Street names are standardized (e.g., converting 'Main St' to 'Main Street'), directionals are normalized (e.g., 'N' to 'North'), and formatting conventions are unified. This parameter transformation maintains adaptability to various input formats while improving reliability by creating consistent, standardized representations for matching.
Solution Approach 2:
The system introduces intermediary representation layers between the diverse input formats and the reference map data. Through decomposition into components and purification of each component, the system creates a standardized intermediate representation that can reliably match with reference data regardless of the original input format variations.
3Adaptability or versatility
If abbreviations and variant names are used to match user input, then the system becomes more flexible, but the matching precision deteriorates due to ambiguous or incorrect associations
Solution Approach 1:
The address is segmented into discrete components (street number, street name, directionals, postal code, city, state, zip), and each component is processed independently. This segmentation allows the system to maintain precision by matching each component accurately while still accommodating abbreviations and variants within each component's processing.
Solution Approach 2:
The system uses feedback mechanisms where the decomposition and purification results are fed back into the composition stage to generate multiple candidate matches. The matching process iterates, using feedback from initial matches to refine subsequent attempts, ensuring high precision by validating and correcting abbreviations and variants through iterative refinement.
Data Source
AI summary
A geocoding method including receiving input location information (212) for a particular location, decomposing (210) and purifying (220) the input information before decomposing (230) the information, and generating (240) multiple candidate locations based on components of the decomposed information, and then normalizing the candidate locations before comparing the candidate locations with information in a location information database.