Research And Implementation Of Recommendation System For Technological Project Assessors

With the explosive growth of internet information resources, the traditional information retrieval technique that based on keyword search can’t meet people’s increasing requirements. Personalized recommendation system has appeared in this case, and it has been widely used in many fields such as E-Commerce recommendation, film and television recommendation, article recommendation, etc. However, there are rarely research and application about evaluation expert automatic recommendation system in the field of technical project assessment. Based on the abundant library information of specialists, it is necessary to automatically recommend appropriate evaluation experts according to different technical projects, which is a guarantee of quality and effectiveness for project review. This paper studies the technical projects assessment recommendation system according to evaluation criteria of technical projects, expert information and technology projects information, and achieves selection of technical project evaluation experts efficiently and accurately. The contents of this paper are summarized as follows:(1) Proposes a technical project evaluation experts recommendation system model. The model builds the projects and evaluation expert vector space model, and makes similarity matching with content-based recommendation algorithm to produce the preliminary list of recommended experts; then adjusts the experts list with collaborative filtering algorithm; at last it generates the final experts list taking expert’s grading as weighting factor.(2) Develops the technical projects and evaluation expert vector space model. It gets key words by projects text segmentation, then calculates the weight of them with TF-IDF algorithm and obtains the key word as well as the corresponding weight which is larger than the threshold, finally it builds the technical projects vector space model based on vector space method. Similarly it is almost the same way to build evaluation experts vector space model.(3) Builds the experts evaluation mathematical model. It builds the model with the information of experts, research projects, articles, personnel training and awarding and so on, then makes adjustment on the recommendation results setting experts grading got from the model as weighting factor.(4) Proposes a hybrid recommendation algorithm of technical projects evaluation experts. Based on the problem of different indistinguishable field has different influence to recommendation with traditional content-based recommendation algorithm, it proposes the content-based algorithm at different levels, and produces the preliminary list of recommended experts, then makes filtering with collaborative filtering algorithm and determines the final evaluation experts list with the help of weighting factor’s hybrid recommendation.The system proposed by this paper has been put into use in ZheJiang technical projects management system, and practical application shows that the recommendation system is feasible and effective.