Research On The Recommendation System For Office Automation System In Colleges

In Chinese major universities, the establishment, operation and the maintenance ofthe college-level office systems is of great significance for the commencement ofnormal teaching and scientific research and management for the entire college. With therising number of students and the diversification of educational nature, the data andinformation need to be addressed in the office is more and more complicated. It is moreand more difficult to complete the desired task with manual configuration method.Therefore, how to properly and efficiently configure and optimize the teachingresources is a key technology in the office system. In order to optimize the allocation ofteaching resources, we can design an automated recommendation system to provideusers with the appropriate required resources. In this context, the paper designed andimplemented a system used to recommend document information personally. Accordingto the specific needs of different users, we will use a set of effective solutions toautomatic classify the collected information and evaluate the content which the usershave never red in order to achieve the personalized recommendation and distribution ofinformation resources. The main content of this paper:1, Research the tokenization method of document information, on the basis ofanalyzing and comparing the advantages and disadvantages with existing methods,generate directed acyclic graph on the basis of the existing dictionary. Use dynamicprogramming method to find the best segmentation way as the system output. Thismethod has a good performance and can relatively accurate segmentation which is notin the dictionary.2, Discussed the method of calculating the similarity between resources and users.Through the establishment of vector space model we analysis the attributes of user andresource, put forward a set of effective algorithms to make the users and resources to bea series of vectors in the space. The degree of similarity between the resources and theusers is simulated by calculating the similarity between the user vector and resource vector.3, We have discussed how to recommend the information personal. We haveintroduced the basic information and the current state of development of recommendedsystem. We also introduced the structure and kinds of main technology of the system.Based on the demand of office in college, we combined the advantages anddisadvantages of current recommendation system and discussed a hybridrecommendation algorithm which is suitable for solving the current problems. Then, wegive the implementation steps of the algorithm and experimental results, finally weanalysis the accuracy of the recommendation algorithm.