机器学习算法在短期电力负荷预测中的应用

为了提高短期电力负荷预测准确性,提出一种蚁群算法(ACO)优化支持向量机(SVM)的短期电力负荷预测模型(ACO-SVM)。首先采用混沌理论对短期电力负荷样本进行重构,然后将SVM参数作为蚂蚁的位置向量,通过蚁群信息交流和相互协作找到SVM最优参数,最后建立短期电力负荷的最优预测模型,并采用实际短期电力负荷数据进行有效性验证。结果表明,ACO-SVM能够准确刻画短期电力负荷变化特性,提高了短期电力负荷的预测准确性。

Application of machine learning in Short-term load prediction

In order to improve the prediction accuracy of short-term load, this paper proposes a short-term load prediction model based on support vector machine and ant colony optimization algorithm (ACO-SVM). Firstly, the data of short-term load are reconstructed by chaotic theory, and then the parameters of SVM were considered the position vector of ants, the optimal parameters are got by ACO, lastly, the optimal prediction model of short-term load is built and the performance of model is tested. The results show that ACO-SVM can describe the change rule of short-term load accurately and improves prediction accuracy.