论文部分内容阅读
针对公共建筑基线负荷难以有效预测的问题,提出了一种基于模糊C-均值聚类预处理的人工神经网络预测方法。采用聚类算法,将大量的复杂历史数据集划分成多个群体的混合,每个群体对应单独的预测模型进行预测。该方法减少了培训数据,克服了标准方法数据量大和处理速度慢的缺点。将预测结果与标准的人工神经网络方法相比较,得到了较高的预测精度,能有效预测公共建筑基线负荷。
Aiming at the problem that it is difficult to effectively predict the baseline load of public buildings, an artificial neural network prediction method based on fuzzy C-means clustering preprocessing is proposed. Using the clustering algorithm, a large number of complex historical data sets are divided into a mixture of multiple groups, each group corresponding to a separate prediction model to predict. This method reduces the training data and overcomes the shortcomings of the standard method such as large amount of data and slow processing speed. Comparing the prediction results with the standard artificial neural network method, a higher prediction accuracy is obtained, which can effectively predict the baseline load of public buildings.