论文部分内容阅读
针对机电设备运行状态受多因素影响,变化趋势复杂,难以用单一预测方法进行有效预测的问题,提出一种新的基于改进灰色系统—支持向量机—神经模糊系统的智能混合预测模型。该模型首先利用改进灰色系统弱化数据序列波动性、支持向量机处理小样本和模糊神经系统处理非线性模糊信息的优点,分别进行趋势预测,然后通过改进遗传算法对这三者的预测结果进行自适应加权组合。将该模型应用于信号随机波动性较强、趋势变化复杂的标准算例和某机组振动趋势的预测中,研究结果表明,该模型的预测性能均优于上述三种单一预测方法。
Aiming at the problem that the operation status of electromechanical equipment is affected by many factors and the trend of change is complex and it is difficult to predict effectively with single prediction method, a new intelligent hybrid forecasting model based on improved gray system-support vector machine-neuro-fuzzy system is proposed. The model firstly uses the improved gray system to weaken the volatility of the data series, the support vector machine to deal with the small sample and the fuzzy neural system to deal with the nonlinear fuzzy information. Then the trend prediction is carried out respectively. Then, the improved genetic algorithm Adapt to weighted combinations. The model is applied to the standard case with strong random fluctuation and complicated trend trend and the prediction of the vibration tendency of a certain unit. The results show that the prediction performance of this model is better than the above three single prediction methods.