A Neural Network-based Research Performance Service Portfolio Evaluation Model and Its Implementation

Nowadays, with the reform of the research evaluation mechanism proposed by the nation, the evaluation mechanism of scientific research activities has been developed rapidly in universities which is strongly supported by the nation. Scientific research evaluation affects the development direction of scientific research activities, thus, how to carry out effective scientific research evaluation activities becomes an important issue among the universities. In this study, a neural network-based service evaluation model is proposed for the existing problems in scientific research evaluation service. The model is divided into two parts: obtaining evaluation index weights based on neural networks and obtaining final evaluation results based on the multi-model fusion method. Finally, the experiments on real university research data are conducted. The results show that our model can effectively obtain research evaluation results, which include both faculty evaluation results and team evaluation results.