Dynamic Vocabulary Spell Checking for User-Specific Terms

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional spell checkers often fail to provide relevant corrections for user-specific terms like neologisms, acronyms, and slang, as they are not included in the initial vocabulary, and even after adding custom words, the proposed corrections may not be relevant due to the comparison criteria used.

Innovation Solution

A computer-implemented method for spell checking that provides a user interface for managing files, where words are selected based on both Levenshtein distance and occurrence data within the user's files, prioritizing words that are frequently used by the user, and updating the vocabulary dynamically to reflect changing user habits.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If a spell checker uses a fixed initial vocabulary for correction suggestions, then the system structure remains simple, but it cannot provide relevant corrections for user-specific terms like neologisms, acronyms, and slang

Engineering Contradiction:
ImproveAbility to handle user-specific termsVSAvoidVocabulary management complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent implements a dynamic vocabulary system that evolves from a fixed initial vocabulary to a flexible structure incorporating user-specific terms. The spell checker automatically learns and adapts to user lexicon by analyzing user files and correcting terms, transforming the static vocabulary into a dynamic one that grows with user needs.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system performs preliminary analysis of user files to build a customized vocabulary before actual spell checking begins. By pre-processing user documents and extracting domain-specific terms, the spell checker is prepared to recognize and correct user-specific terminology rather than treating all unknown words as errors.

Inventive Principle:
Principle #10Preliminary action

2Measurement precision

If a spell checker uses Levenshtein distance as the primary correction criterion, then the correction algorithm remains simple, but it proposes irrelevant corrections for user-specific words even after adding them to the vocabulary

Engineering Contradiction:
ImproveCorrection relevance accuracyVSAvoidCorrection algorithm complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent changes the selection parameter from purely Levenshtein distance to a hybrid criterion combining distance with frequency of occurrence in user files. This parameter transformation allows the system to prioritize corrections based on both textual similarity and contextual relevance, significantly improving correction accuracy for user-specific terms.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The system incorporates feedback from user file analysis into the correction selection process. By monitoring which corrected terms actually appear in user documents, the spell checker learns and adjusts its correction priorities, creating a feedback loop that continuously improves correction relevance.

Inventive Principle:
Principle #23Feedback

3Measurement precision

If a spell checker analyzes all user files to build a customized vocabulary, then the personalization accuracy improves, but the time and computational resources required increase

Engineering Contradiction:
ImproveUser lexicon matching accuracyVSAvoidVocabulary building time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent applies partial action by analyzing only the most relevant portions of user files rather than processing every document in exhaustive detail. The system strategically samples user content to build a representative vocabulary, achieving high personalization accuracy without the full time cost of complete file analysis.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP1855210B1Spell checking
Publication Date: 2018.01.03 DASSAULT SYSTEMES SA
  • EP1855210B1 patent drawingFigure 1
  • EP1855210B1 patent drawingFigure 2
  • EP1855210B1 patent drawing

AI summary

The invention concerns a computer-implemented method for spell checking, comprising the steps of: - providing a user with a user interface adapted for managing files stored on at least one computer; - receiving (S30) via the user interface a user action involving at least one first word; - returning (S80) via the user interface at least one second word selected according to: - a distance from said at least one second word to said at least one first word; and - data of occurrence of said at least one second word in said files. The invention is also directed to a computer program product and computer system allowing for taking steps according to the invention.