摘要： 对非标记化表格进行数据建模,利用非标记化表格在文档中的结构分布特征,给出了非标记化表格的抽取算法。对非标记化表格进行行列划分,然后进行标题归纳和单元格合并。实验结果表明,论文提出的算法的正确性令人满意。

关键词: 非标记化表格, 信息抽取, 分层聚类

Abstract: Based on the data modeling of the untagged table,this paper proposes an extraction algorithm by using its structural distribution features in documents.It splits the untagged table into rows and columns,and then inducts headers and merges cells.Experimental results indicate that the accuracy of the algorithm is satisfactory.

中图分类号: 

• TP311.11