异构专利数据源集成方案设计与实现

针对目前用于专利分析的数据存在来源单一、预处理操作不够、可挖掘程度浅等问题,设计并实现异构专利数据源集成方案,即从七国两组织的专利数据库获取数据到本地专利数据库;以本地数据库为基础数据源,利用SSIS工具通过ETL(数据抽取-数据转换-数据装载)操作,生成规范的、集成的高质量数据;进而将其加载到事先围绕KPI(关键性能指标)分析构建好的专利数据仓库中,从而为专利多维分析以及数据挖掘提供有效的数据支持。

Design and Implementation of Data Integration over Heterogeneous Patent Sources

With consideration of the problems concerning the data of patent analysis, such as single data source, rough pretreatment, and low-level data mining, this paper designs and achieves the data integration over heterogeneous patent sources. Specifically, the local patent database where the data are acquired from heterogeneous sources including two organizations and seven countries is regarded as basic data source. After using the SSIS tool for data cleaning and data transformation, the data from local database are loaded into data warehouse that is built according to the key performance indicators, which provides data support for more advantaged analysis.