摘要
针对当前创新绩效评价模型存在收敛效率低以及泛化性能弱的缺陷,提出基于BP神经网络的创新绩效评价模型。对创新绩效评价的指标体系进行构建,并收集相应的创新绩效评价数据,采用BP神经网络模拟人脑对创新绩效数据进行训练,并采用梯度法确定BP神经网络的参数,建立创新绩效评价模型,最后通过仿真实验测试其性能。实验结果表明,该模型提高了创新绩效评价的精度,而且评价速度得到大幅度提高,评价效果明显优于其他模型,具有更高的实际应用价值。
Since the available innovation performance evaluation model has the defects of low convergence efficiency and weak generalization performance, an innovation performance evaluation model based on BP neural network is proposed. The index system of the innovation performance evaluation was constructed. The corresponding innovation performance evaluation data is collected, and trained with BP neural network simulating the human brain. The gradient method is used to determine the parameter of BP neural network. The innovation performance evaluation model was established, and its performance was tested with simulation experiment. The experimental results show that the model can improve the accuracy of innovation performance evaluation and evaluation speed, its evaluation effect is superior to other models, and has high practical application value.
出处
《现代电子技术》
北大核心
2017年第15期56-58,63,共4页
Modern Electronics Technique
关键词
BP神经网络
评价模型
收敛效率
泛化性能
BP neural network
evaluation model
convergence efficiency
generalization performance
