摘要
针对传统机器学习算法的泛化性能不足,模型参数与结构确定困难等缺点,采用基于灰色关联和麻雀搜索算法(SSA)的组合算法,优化最小二乘支持向量机(LSSVM)参数.应用投影原理改进传统灰色关联相似日选取算法,采用SSA对LSSVM进行寻优.以某地区的负荷数据为例,进行短期电力负荷预测.研究结果表明:采用麻雀搜索算法优化的LSSVM预测精度更高,稳定性更好.SSA优化的LSSVM相对误差平均值比LSSVM和粒子群优化算法(PSO)-LSSVM分别减少2.96%和0.95%,平均相对误差分别减少2.58%和1.46%,均方根相对误差分别减少2.71%和1.46%.
In order to overcome the shortcomings of traditional machine learning algorithms,such as lack of generalization performance and difficulty in determining model parameters and structure,a combined algorithm based on grey correlation and sparrow search algorithm(SSA)was proposed to optimize the parameters of least square support vector machine.The projection principle is used to improve the traditional grey correlation similarity day selection algorithm,and the sparrows search algorithm is used to optimize the parameters of LSSVM.The short-term load forecasting based on the load data of a certain area in south China is compared with Lssvm and particle swarm optimization algorithm(PSO)-LSSVM.The results show that LSSVM optimized by sparrow search algorithm is more accurate,more accurate and more stable than traditional LSSVM and LSSVM optimized by PSO.
作者
张子阳
王珂珂
ZHANG Ziyang;WANG Keke(Faculty of Electrical and Control Engineering,Liaoning Technical University,Huludao 125100,China)
出处
《辽宁工程技术大学学报（自然科学版）》
CAS
北大核心
2022年第3期283-288,共6页
Journal of Liaoning Technical University (Natural Science)
作者简介
张子阳(1993-),男,黑龙江黑河人,硕士,助理工程师,主要从事电力传动技术与应用方向的研究.
