基于深度神经网络的电力工程数据聚类模型设计

Design of power engineering data clustering model based on deep neural network

针对实际应用中电力工程数据的维度不断升高,而现有聚类模型在处理高维数据时模型有效性降低的问题,提出一种基于深度神经网络的电力工程数据聚类模型。该模型在对原始高维数据进行预处理的基础上,引入了深度受限玻尔兹曼机神经网络。通过深度神经网络的无监督学习,将高维的原始电力工程数据进行非线性化处理,从而映射到低维空间。在低维空间采用模糊聚类算法对数据进行处理,得到对原始数据的有效聚类分析。采用实际电力工程数据对模型进行仿真验证的结果表明,该模型可以明显提高电力工程数据聚类的有效性。

Aiming at the problem that the dimension of power engineering data is increasing in practical application,and the effectiveness of the existing clustering model is reduced when dealing with high⁃dimensional data,a power engineering data clustering model based on deep neural network is proposed.On the basis of preprocessing the original high⁃dimensional data,the model introduces the depth limited Boltzmann machine neural network.Through the unsupervised learning of the depth neural network,the high⁃dimensional original power engineering data are nonlinear processed,and then mapped to the low dimensional space.In the low dimensional space,the fuzzy clustering algorithm is used to process the data,and the effective clustering analysis of the original data is obtained.The simulation results show that the model can significantly improve the effectiveness of power engineering data clustering.