科技项目查重是避免重复立项、重复建设的重要手段，然而目前仍然缺乏行之有效的方法。本项目拟采用大数据挖掘理论和方法对与项目相关的多来源信息进行整合，构建准确、稳定的项目相似度判别模型，实现准确、高效的项目查重方法。围绕科技项目的大数据网络是项目查重的数据基础，将分别构建项目信息数据库，文献题录数据库，负责人数据库，承担单位数据库，以及项目与关键词的关联数据库。基于上述理论和数据基础，结合Hadoop技术，实现用于电动汽车领域项目查重的原型实验系统，为后续正式平台的搭建和推广积累经验和奠定基础。本研究工作为解决项目查重问题提供了一种全新的思路和方法，研究结果将是对现有项目查重方法的有益扩展和补充。

Similar project detection is an important way to avoid repeated project and constuction for project management department. however, how to effectively find similar projects is still an unresolved problem. This project will use big data mining theory and technology to integrate multi-source information of a project, construct a new discriminative model of project similarity, and hereby achieve a similar project detection method effectively. Scientific project information database is the foundation of implementing the above detection method. Five databases will be constructed, which is project information database, literature information database, experts database, institution database, and project-literature associated information database. Base on above database and theory, we will implement a prototype system for the field of electric vehicles using Hadoop software architecture. This project will provide an novel method of detecting similar projects, and the results will extend and improve the existed theory and method.