Input Method Candidate Segmentation for Faster Word Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing input methods require users to manually correct multiple incorrect words in entered text, leading to increased modification workload due to the lack of accurate candidate terms in the first-page candidate term set.
Innovation Solution
An input method that detects pre-modification operations and displays a first-page candidate term set including both original and sub-candidate terms with higher likelihood of being correct, based on user input, to facilitate quick and accurate word corrections.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If the input method application displays only traditional candidate terms matching the full input sequence, then the candidate term set is simple to generate, but the user cannot quickly find and correct incorrect words among multiple entered words
Solution Approach 1:
The patent segments the input sequence into multiple individual words and generates candidate terms for each word separately. When a user inputs multiple words (e.g., pinyin for multiple characters), the system divides them into discrete units, generates candidate terms for each unit independently, and displays them in the candidate term set. This segmentation allows users to quickly identify and correct incorrect words among the segmented candidates rather than searching through a single undifferentiated candidate list.
Solution Approach 2:
The patent introduces a new dimension to candidate term generation by creating sub-candidate terms that correspond to individual words within the input sequence, in addition to the traditional original candidate terms that match the full sequence. This dimensional expansion of the candidate term structure enables users to operate at different levels (full sequence vs. individual word) to efficiently locate and correct errors.
2Reliability
If the user manually corrects each incorrect word one by one in large text inputs, then the correction process is thorough, but the modification time and workload increase significantly
Solution Approach 1:
The patent performs preliminary segmentation of the input sequence into individual words and generates candidate terms for each word before the user needs to correct errors. By pre-processing the input in this way, the system prepares the candidate term set in advance, allowing users to quickly scan and identify incorrect words without manual step-by-step correction, thus reducing modification time while maintaining correction accuracy.
Solution Approach 2:
The system provides feedback to the user by displaying segmented candidate terms corresponding to each word in the input sequence. This feedback mechanism allows users to immediately see alternative interpretations for each word segment, enabling quick identification and correction of incorrect words without time-consuming manual verification.
3Productivity
If the first-page candidate term set includes only original candidate terms matching the complete input, then the candidate generation logic is simple, but relevant candidate terms for individual words may not appear on the first page
Solution Approach 1:
The candidate term set is segmented into two types: original candidate terms matching the full input sequence and sub-candidate terms matching individual words. This segmentation ensures that relevant candidate terms for individual words appear prominently in the first-page display, improving modification efficiency without requiring complex multi-page navigation.
Solution Approach 2:
The patent merges original candidate terms and sub-candidate terms into a unified first-page candidate term set. This combination allows both types of candidate terms to coexist and be displayed together on the first page, providing users with comprehensive options for both complete phrase matching and individual word correction in a single view.
Data Source
AI summary
In an input method, an electronic device detects a pre-modification operation on M target words. The pre-modification operation includes entering first input information that corresponds to at least N words, where N is greater than M. The electronic device displays a first-page candidate term set in an interface in response to the pre-modification operation. The first-page candidate term set includes at least one original candidate term corresponding to the first input information and at least one sub-candidate term corresponding to some of the first input information. The original candidate term includes at least N words, and the sub-candidate term includes M words.