Methods Of Project Evaluation Experts Recommend

Expert recommendation method for project evaluation can recommend suitable evaluation experts automatically in accordance with the declaration information of the project, it can provide valuable information for the competent bodies selection evaluation experts. The core idea of evaluation experts recommendation is using the information of the description documents of both declaration project and candidate experts, then obtaining the correlation of the project and the candidate experts, and also combining the expert’s experience in project evaluation with his operational capacity, scoring and sorting the candidates to pick out the evaluation experts with high quality. Focused on the characteristics of recommending evaluation experts automatically, this paper carry out exploration and research around the problems of obtaining the topic information form the documents of project and expert, measuring the correlation of the project and expert, and designing the selection method for expert recommendation. We had accomplished innovative achievements as follows:(1) This paper discussed a method of obtaining topic information from the description documents of the projects and the experts. Firstly, the method analyzed the attributive characteristics of the documents. Then, processed the attributive columns of the documents according to their forms of representation:the columns that concise and characterized by structured information would only take operations of word segmentation and stop words filtering, as for the columns that characterized by lots of text, we used LDA (Latent Dirichlet Allocation) topic model to obtain topic words from them. Finally, the method laid the foundation for using the information of documents effectively by characterizing the documents with the topic words.(2) This paper presented an expert recommendation method based on the topic information. Firstly, the method constructed the topic feature space of the documents though the method of statistical topic words frequency. Secondly, combined the importance factors of the document columns, we made use of TFIDF method to build topic feature vectors of both project documents and expert documents respectively. Finally, improved algorithm of similarity calculation is used to compute the correlation of the project and each expert, the experts with a high correlation of the project are chosen as the result of recommendation. Experiment show that the proposed method has a good effect of recommending experts for project evaluation.(3) This paper designed an expert selection method for expert recommendation that synthesized the correlation of the project and the expert, expert’s experience in project evaluation and the operational capacity of the expert. After obtaining the correlation of the project and each expert, this method firstly found out the collection of evaluation experts that corresponding to the historical projects which similar with the new project from the historical assessment data, by calculation the similarity between the projects, and then got their experience in projects evaluation. Secondly, we used the information in the expert document to evaluate the operational capacity of each expert from four aspects as follows:basic qualifications, academic achievements, participating in projects and received awards. Finally, we rated and sorted candidate experts by synthesizing three factors above, to obtain the final recommended experts.(4) This paper designed and implemented an expert recommendation prototype system for project evaluation, this prototype system could provide convenience for the further research of expert recommendation method.