关系数据库SQL代码的自动评分算法研究  

Research on Automatic Scoring Algorithm for SQL Code in Relational Database

作　　者：吴娇娇 何小卫[1] 赵洋 WU Jiao-jiao;HE Xiao-wei;ZHAO Yang(School of Mathematics,Physics and Information Engineering,Zhejiang Normal University,Jinhua 321004,China)

机构地区：[1]浙江师范大学数理与信息工程学院,浙江金华321004

出　　处：《计算机技术与发展》2018年第4期114-118,共5页Computer Technology and Development

基　　金：国家自然科学基金(61572023);浙江省自然科学基金(LY14F010008)

摘　　要：针对数据库教学中SQL(结构化查询语言)代码评判的不规范性及不明确性,建立了一个SQL代码自动评分模型。该模型首先对SQL代码进行了预处理,分析语义结构进而提取出代码的特征,同时进行分词处理;然后构建SQL代码的同义库,根据同义库将分词后的代码片段进行同义转换;之后运用最长公共子串算法对代码片段进行相似度匹配,进而赋予各个片段影响因子,得到代码总的相似度;最后分析人工评分数据制定相应的自动评分策略,从而评定出SQL代码的分数。该模型首次构建同义库,有效地提高了SQL代码的评分效率,可以达到比较准确的水平,采用循环匹配也可以对SQL中的长代码进行评分。In view of the uncertainty and ambiguity of SQL code evaluation in database teaching,we establish an automatic scoring model for SQL code.Firstly,this model preprocesses the SQL code,analyzes the semantic structure as to extract the characteristics of the code,and executes the word segmentation.Secondly,we build a synonymous library of SQL code and make the code snippet synonymy conversion after the word segmentation.And then the model uses the longest common substring algorithm to calculate the similarity of code fragments,and gives the impact factor of each fragment as to get the code total similarity.Finally,we analyze the manual scoring data to develop the corresponding automatic scoring strategy to assess the scores of the SQL code.The technique proposed for the first time builds a synonymy library which effectively improves the efficiency of SQL code scoring and achieves a more accurate level.The model can also be used to rate long code in SQL with the use of circular matching.

关 键 词：代码自动评分 结构化查询语言 最长公共子串 同义库 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]