[1]魏自强,班元郎,徐 伟,等.短文本聚合在元器件供方匹配中的应用与研究[J].计算机技术与发展,2022,32(07):216-220.[doi:10. 3969 / j. issn. 1673-629X. 2022. 07. 037]
WEI Zi-qiang,BAN Yuan-lang,XU Wei,et al.Application and Research of Short Text Aggregation in Component Supplier Matching[J].,2022,32(07):216-220.[doi:10. 3969 / j. issn. 1673-629X. 2022. 07. 037]
点击复制
短文本聚合在元器件供方匹配中的应用与研究()
《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]
- 卷:
-
32
- 期数:
-
2022年07期
- 页码:
-
216-220
- 栏目:
-
应用前沿与综合
- 出版日期:
-
2022-07-10
文章信息/Info
- Title:
-
Application and Research of Short Text Aggregation in Component Supplier Matching
- 文章编号:
-
1673-629X(2022)07-0216-05
- 作者:
-
魏自强; 班元郎; 徐 伟; 王文玺
-
贵州航天计量测试技术研究所,贵州 贵阳 550009
- Author(s):
-
WEI Zi-qiang; BAN Yuan-lang; XU Wei; WANG Wen-xi
-
Guizhou Aerospace Metrology and Testing Technology Research Institute,Guiyang 550009,China
-
- 关键词:
-
Jaro-Winkler 算法; Levenshtein 距离; 短文本聚合模型; 数据特征; 供方匹配
- Keywords:
-
Jaro-Winkler algorithm; Levenshtein distance; short text aggregation model; data characteristics; supplier matching
- 分类号:
-
TP315
- DOI:
-
10. 3969 / j. issn. 1673-629X. 2022. 07. 037
- 摘要:
-
航天采购部门采购合格供方的元器件是保证航天用元器件可靠性的方法之一。 确定供方是否在合格供方目录中,是航天元器件采购流程中的一个重要步骤。 但由于航天各院所系统中对供方定义标准不一致,常以供方公司的别称、简称代替供方名称,这导致同一供方出现多种不同名称,这给如何匹配合格供方带来了挑战。 针对航天各院所系统中的供方数据的特征,提出了一种结合 Jaro-Winkle 算法和 Levenshtein 算法的融合算法。 该算法通过引入调整阈值及系数,将字符的位序、字符替换、添加、删除操作等因素纳入到供方名称的短文本相似度计算中,提高供方名称的短文本匹配准确率。 通过在航天元器件合格供方匹配流程中的应用,该算法有效提高了供方的匹配准确率。
- Abstract:
-
The procurement of components from qualified suppliers by aerospace procurement department is one of the methods? ? to ensure the reliability of aerospace components. Determining whether the supplier is in the list of qualified suppliers is? an important step in the procurement process of aerospace components. However, due to the inconsistency of supplier definition standards in the systems of aerospace institutes,the supplier’s nickname and abbreviation are often used to replace the supplier’s name, leading to a variety of different names for the same supplier,which brings challenges to how to match qualified suppliers. According to the characteristics of supplier data in the systems of aerospace institutes,we propose a fusion algorithm combining Jaro-Winkle algorithm and Levenshtein algorithm. By introducing the adjustment threshold and coefficient,the algorithm integrates the character bit order,character replacement,addition,deletion and other factors into the short text similarity calculation of the supplier’s name,so as to improve the short text matching accuracy of the supplier’s name. Through the application in the qualified supplier matching process of aerospace components, the proposed algorithm can effectively improve the matching accuracy of suppliers.
更新日期/Last Update:
2022-07-10