Research And Realization On Document Repository Based On DSpace

With the improvement of teaching and research work in universities, the science and technology document increased gradually at the same time. Many colleges and universities have accumulated plenty of valuable academic information including electronic documents, pictures, books, and experiment data, etc. How to manage these data is a challenge to every university.To preserve the digital resources for a long time and to utilize the present information of knowledge and document comprehensively and effectively and to realize a better free communication among academic achievements during teaching and scientific researches, the thesis analyzed and compared some software for establishing open access repository at home and abroad and found that DSpace is the best open-source architecture for digital storage. It has found that DSpace can be used to gain, save and release all kinds of digital resources and it conforms to our assumption of saving data for a long time. Meanwhile, the functions of the system can be customized to meet the practical needs of Inner Mongolia Agricultural University.The thesis researched and realized a document and information resource repository system based on DSpace, which was localized and its function was extended. The paper also analyzed and compared the Lucene retrieval toolkit of DSpace and made it support full text retrieval in both Chinese and English. Through comparative experiments on many normal Chinese word segmentation methods, the paper has determined to use JE-analyzer as Chinese word segmentation methods in the retrieval toolkit and has constructed synonym dictionary to increase the recall ratio by query expansion.