A Distributed Exam Item Bank System Based on Hadoop Ecosystem

Exam item bank system is an important precondition for the separation of teaching and examination and helpful for easing the burden of teachers. It is also the foundation of assessing teaching achievements equitably and objectively, being useful for supervising teachers teaching. Therefore, this paper designed a distributed item bank system with high scalability based on a popular big data processing platform, Hadoop. The system exploited Hadoop distributed file system for unstructured item data storage, and utilized Hadoop ecosystem to respectively design functions of test paper generation, test paper search, item retrieval, test result check, test result analysis, item similarity retrieval, and so on. The paper presented the formalization description of the test paper generation problem, based on which users could design smart and efficient test paper generation algorithms based on the methods of heuristic, metaheuristic, machine learning, or any others in the system.