基于关联度的混沌序列局域加权线性回归预测法

分析了基于欧氏距离局域预测法存在的缺点，在此基础上提出了一种基于关联度的局域加权线性回归预测法。该方法以关联度代替欧氏距离作为判别不同相点间相关性的准则，并将相点间的相关性大小通过“加权”的方式作用于混沌序列预测模型，从而克服了局域线性回归预测法的缺点。首先对新方法的原理及其合理性进行了系统阐述；然后推导了其算法过程；最后将该方法应用于电力系统短期负荷的预测中，得到了理想的预测结果。通过分析和比较，验证了所提方法的有效性。

LOCAL ADDING-WEIGHT LINEAR REGRESSION FORECASTING METHOD OF CHAOTIC SERIES BASED ON DEGREE OF INCIDENCE

The defect of local forecasting method based on Euclidean distance is analyzed. On this basis, a novel method called local adding-weight linear regression forecasting method based on degree of incidence is proposed. In this method, degree of incidence, instead of Euclidean distance, is used as criterion to judge the correlation between different phase points. At the same time, the values of expressing correlation are acted on chaotic series forecasting model by means of adding-weight. It overcomes the defect of local linear regression forecasting method. Firstly, the principle and reasonability of the new method are demonstrated systematically. Then, its algorithm process is derived. In the end, this method is applied to the short-term load forecasting of power system, and get ideal results. Through analyzing and comparing, the validity of the suggested method is verified.