在用混沌理论和神经网络进行短期负荷预测时，神经网络的输入的选择至关重要，该程序用matlabl实现了基于混沌时间序列的嵌入维数的选择-using chaos theory and neural networks for short-term load forecasts, the neural network is essential to choose an input, The procedure used matlabl achieved a chaotic time series bas

采用KL变换的G-P 方法计算混沌吸引子关联维，计算时间序列的最优嵌入维数。(KL transform g-p method is used to calculate the correlation dimension of chaotic attractor and the optimal embedding dimension of time series.)