Syntax Parsing Apparatus Using Preprocessing and Statistical Disambiguation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing syntax parsing methods face challenges in reducing ambiguity and increasing processing efficiency, particularly in rule-based parsing, where complexity and error propagation are issues, and statistics-based methods struggle with data deficiency and slow parsing speeds.
Innovation Solution
A syntax parsing apparatus and method that incorporates preprocessing techniques, including pattern chunk recognition, basic noun phrase statistical modeling, and syntax tree generation, to resolve ambiguity and enhance parsing correctness and efficiency by applying statistical models learned from a syntax tree attachment corpus.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If rule-based syntax parsing is applied, then parsing structure is clear and controllable, but ambiguity processing is limited and parsing complexity increases
Solution Approach 1:
The patent segments the syntax parsing process into two distinct phases: a rule-based parsing phase that handles clear syntactic structures, and a statistics-based disambiguation phase that resolves ambiguous cases. This segmentation allows each phase to specialize in what it does best, reducing overall complexity while maintaining reliability.
Solution Approach 2:
The patent introduces statistical models as an intermediary layer between rule-based parsing and final syntax tree generation. This intermediary handles the ambiguous cases that rule-based parsing cannot resolve, using statistical information from corpora to make informed decisions about syntactic structure.
2Reliability
If statistics-based syntax parsing is applied, then ambiguity can be resolved statistically, but parsing speed decreases due to search of massive statistic parameter space
Solution Approach 1:
The patent performs preliminary rule-based parsing to establish a initial syntactic structure before applying statistical disambiguation. This preliminary action reduces the search space for statistical models, as they only need to resolve ambiguity in already-partially-structured data rather than parsing from scratch.
Solution Approach 2:
The patent applies statistical models selectively only to ambiguous portions of the syntax parsing problem, rather than applying them uniformly to all parsing decisions. This local application of statistical methods maintains high parsing speed while resolving ambiguity where it actually occurs.
3Reliability
If general statistics-based syntax parsing is applied, then ambiguity can be resolved, but learning data deficiency decreases correctness and processing efficiency
Solution Approach 1:
The patent uses universal statistical linguistic features and patterns that can be learned from general language corpora and applied across different domains and languages. This multi-functionality allows the statistical models to achieve good performance even with limited domain-specific learning data.
Data Source
AI summary
The present disclosure relates to a syntax parsing apparatus based on syntax preprocessing and a method thereof. In specific, the present disclosure parses syntaxes that can be parsed by rules and patterns without ambiguity by syntax parsing preprocessing, draws all possible syntax parsing results by applying syntax rules based on a result of syntax parsing preprocessing in which ambiguity is partially resolved, and resolves structural ambiguity by applying a statistic syntax parsing model learned from a syntax tree attachment learning corpus so as to reduce ambiguity in rule-based syntax parsing and to resolve ambiguity by a statistics-based scheme so that parsing correctness and processing efficiency in a syntax parsing method can be enhanced.