基于迭代匹配和未登录词识别的航空领域文本分词方法
By constructing a relevance matrix and a word formation weight dictionary, and combining semantic similarity and contextual similarity, iterative matching and out-of-vocabulary word recognition are performed, solving the problem of insufficient new word recognition capability and efficiency in Chinese word segmentation in the aviation field, and achieving efficient and accurate word segmentation results.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHINA AERO POLYTECH ESTAB
- Filing Date
- 2023-11-21
- Publication Date
- 2026-07-17
AI Technical Summary
Existing Chinese word segmentation algorithms are insufficient in terms of ability and efficiency in handling new word recognition, especially in the aviation field. Traditional methods rely on dictionaries, have high computational requirements, are prone to word association, and have weak new word recognition capabilities.
By constructing a relevance matrix and a word formation weight dictionary, and combining semantic similarity and contextual similarity, iterative matching and out-of-vocabulary word identification are performed to establish segmentation rules, identify out-of-vocabulary words, and improve word segmentation efficiency.
It effectively alleviates the dependence of traditional word segmentation algorithms on dictionaries, reduces the amount of computation, avoids word concatenation, improves the ability to identify new words and the accuracy of word segmentation, and improves processing efficiency.
Smart Images
Figure CN118070799B_ABST