The University Performance Evaluation with The Decentralization of Local Governmente

摘要

By aiming to the problems of that during the evaluation process on university scientific research capacity, there existed the characters with multiple factors and high non-linearity, and the poor model evaluation accuracy due to the strong subjectivity of the classic evaluation model, in this paper it proposes the evaluation model for university scientific research capacity based on the collaboration of both the intelligent water drop algorithm (IWD) and rough set block neural network (RBNN). First, the IWD algorithm was introduced, and by aiming to the problem that the searching range beside the fixed domain of the traditional IWD algorithm is adverse to promote the efficiency of algorithm search, it proposes the local spatial auto scaling algorithm (LSAS), such an algorithm can automatically adjust the searching space of the next step according to the optimal individual among current colony, give guidance to the evolution; secondarily, based on rough set theory it makes the feature pre-processing on university scientific research capacity data and simplifies the calculated amount of the data; finally it codes the block neural network and rough set parameters and realizes the evaluation on university scientific research capacity. The simulated results show that this evaluation model shall have the higher accuracy and faster calculation efficiency.