Incremental Search Error Compensation on Overloaded Keypads
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current search technologies fail to effectively perform error correction for orthographic and typographic errors, especially in incremental searches on devices with overloaded keys, as they do not handle ambiguous inputs, multiple word queries, or word prefixes and fragments efficiently.
Innovation Solution
A system that uses a database with pre-computed and dynamically generated variants of descriptive terms to map user inputs to likely content items, incorporating error models that account for typographic and orthographic errors, and presents results based on personalized and contextual relevance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If incremental search with reduced text entry is used on devices with overloaded keys, then ease of operation is improved, but measurement precision deteriorates due to ambiguous inputs and errors
Solution Approach 1:
The system pre-computes and stores multiple variants of descriptive terms (including misspellings, phonetic equivalents, and alternative phrasings) in advance in a database. When a user enters an ambiguous or erroneous search query using an overloaded keypad, the system retrieves pre-computed variants that match the input pattern, enabling accurate results without requiring the user to input the complete or correct term.
Solution Approach 2:
The system introduces an intermediary error model and variant matching layer between the user's ambiguous input and the search database. This intermediary component generates and tests multiple possible interpretations of the input (including corrected spellings and alternative phrasings) before presenting results, effectively mediating between imprecise user input and the requirement for accurate search matching.
2Reliability
If error correction methods are applied to complete words, then orthographic error compensation is improved, but adaptability deteriorates for partial prefixes and fragments
Solution Approach 1:
The system applies error correction and variant generation not only to complete words but also to partial prefixes and fragments. When the user enters an incomplete or erroneous prefix, the system generates variants of that partial input and searches for matches, applying the same error compensation techniques used for complete words to partial inputs as well.
Solution Approach 2:
The error correction system is designed to handle multiple types of inputs universally - complete words, partial prefixes, fragments, and multi-word queries - using the same variant generation and matching mechanism. The system adapts the error correction approach based on the input type while maintaining a unified processing framework.
3Measurement precision
If complete word entry is required for accurate search, then measurement precision is improved, but productivity deteriorates due to increased input time and error probability
Solution Approach 1:
The system pre-computes and stores multiple variants of descriptive terms (including misspellings, phonetic equivalents, and alternative phrasings) in advance in a database. When a user enters an ambiguous or erroneous search query using an overloaded keypad, the system retrieves pre-computed variants that match the input pattern, enabling accurate results without requiring the user to input the complete or correct term.
Solution Approach 2:
The system provides feedback to the user by presenting multiple possible matches in order of likelihood when the input is ambiguous or erroneous. This feedback mechanism allows the user to quickly identify and select the intended term without needing to re-input or correct the original query, maintaining both accuracy and efficiency.
Data Source
AI summary
Systems and method provide for finding and presenting content items corresponding to multiple prefixes entered by a user using input logic. The systems and methods enable: storing a plurality of descriptive terms in a database, wherein each descriptive term of the plurality of descriptive terms characterizes a content item; receiving, using input logic, a first set of keystrokes to build a first prefix string; identifying, using mapping logic, a first set of content items corresponding to the first prefix string by comparing the first prefix string to the plurality of descriptive terms; receiving, using input logic, a second set of keystrokes to build a second prefix string; identifying, using mapping logic, a second set of content items corresponding to the second prefix string by comparing the second prefix string to the plurality of descriptive terms; generating, using dynamic intersection logic, a third set of content items appearing in both the first set of content items and the second set of content items to generate a set of most likely content items corresponding to the first and second prefixes; and generating for display, using presentation logic, the third set of content items, wherein the third set of content items is ordered in accordance to an ordering criteria.