Multi-word Candidate Selection for Geolocation Input Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for converting geographical coordinates to memorable word sequences face challenges in efficiently processing and correcting user input errors due to the vast number of possible combinations, limited memory on mobile devices, and the need for computational efficiency.
Innovation Solution
A method that selects candidate words from a fixed dictionary for each part of the input string, using a series of scoring and ranking methods to reduce the number of combinations to be processed, including weighted Levenshtein edit distance and implicit procedures to quickly identify the most likely intended input, with optional steps for morphological variants and reordered variants.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If all possible word combinations are scanned to correct user input errors, then accuracy of correction is improved, but processing time becomes excessively long and computationally infeasible on mobile devices
Solution Approach 1:
The patent segments the vast search space of all possible word combinations into manageable subsets by identifying and processing only those combinations that differ from the input by a small number of edits (e.g., single character substitutions, insertions, or deletions). This segmentation allows the system to focus computational resources on the most likely candidate corrections rather than exhaustively searching all possibilities, thereby achieving accurate corrections within acceptable timeframes on mobile devices.
Solution Approach 2:
The patent applies partial action by performing corrections only for the most probable error types (such as common typos or phonetic similarities) rather than attempting to correct all possible input variations. By limiting the scope of correction attempts to the most likely scenarios based on statistical analysis of user input patterns, the system achieves practical accuracy without the computational burden of exhaustive search.
2Measurement precision
If a comprehensive algorithm is used to scan all possible location identifiers, then correction accuracy is improved, but memory requirements exceed the limited capacity of mobile devices
Solution Approach 1:
The patent extracts and processes only the necessary subset of data required for correction rather than loading all possible location identifiers into memory. By extracting only the relevant candidate combinations that match the input pattern within a small edit distance, the system achieves accurate corrections while keeping memory consumption within the limits of mobile devices.
Solution Approach 2:
The patent performs preliminary filtering of candidate corrections based on simple criteria (such as character distance metrics) before applying more computationally intensive verification algorithms. This preliminary action reduces the volume of data that needs to be held in memory during the correction process, enabling accurate corrections on devices with limited memory resources.
3Reliability
If complex correction algorithms are implemented to handle all possible input errors, then reliability of location identification is improved, but device complexity increases beyond what is practical for mobile implementations
Solution Approach 1:
The patent applies local quality by using different correction strategies for different types of input errors rather than a single uniform complex algorithm. For example, simple typos are corrected using basic character distance metrics, while more complex errors involve phonetic analysis or contextual understanding. This localized approach to error correction maintains reliability for each error type while keeping the overall system complexity manageable for mobile devices.
Data Source
AI summary
Provided is a computer implemented method, device and system for suggesting one or more multi-word candidates based on an input string of characters received at an electronic device. The method comprises the steps of a) for each of a plurality of n selected parts of a received input string, selecting m candidate words from a fixed dictionary of q words, where n, m and q each comprise positive integers, n is greater than or equal to 2 and m is less than q; and b) selecting from the resultant mn combinations of candidate words a subset p of candidate combinations of words where p is a positive integer and p is less than mn. In some embodiments, the multi-word candidates can be resolved to respective unique geographical locations and a subset of the multi-word candidates or their resolved geographical locations can be presented at the electronic device for user selection. It is preferred that the plurality of n selected parts of the received input string comprises x words comprising the received word input string such that n=x.