基于标志词的正向自适应长度匹配的地址分词算法与缺失地址要素补充方法

Address Segmentation Algorithm Based on Forward Adaptive Length Matching by Mark Words and Supplementary Method of Missing Address Elements

目的将电子病历中患者的非结构化地址信息转化成结构化地址信息,并补充地址中缺失的地址要素。方法构建存储标准地址数据集的标准地址库和自定义的地址匹配规则库。依托标准地址库,采用基于地址要素标志的正向自适应匹配算法将地址进行分词。将分词得到的地址要素根据构建的自定义地址匹配规则库从后往前查找,得到完整的地址。结果该方法实现了病历中地址数据的自动分词,同时补充了地址数据中缺失的地址要素,完成地址标准化的工作。结论本研究极大地方便了临床病案首页中地址信息的自动获取、各类机构数据上报和数据统计分析工作,大幅减少人工数据处理的工作量,为后续其他信息的提取和标化打下坚实基础。

Objective To transform the unstructured address information of patients in electronic medical record into structured address information, and supplement the missing address element in the address. Methods A standard address library for storing standard address data sets and a custom address matching rule library were built in this paper. Based on the standard address library, the address was segmented by a forward adaptive matching algorithm based on address elements. Then the address elements obtained by word segmentation were looked up from back to front according to the custom address matching rule base constructed to obtain the complete address. Results The automatic word segmentation of address data in medical records was realized, and the missing address elements in address data was complemented to complete the work of address standardization. Conclusion This study not only greatly facilitates the automatic acquisition of address information on the first page of clinical medical records, but also facilitates the data reporting and statistical analysis of various institutions. It can greatly reduce the workload of manual data processing and lay a solid foundation for subsequent extraction and standardization of other information.