摘要
近年来研究者们提出了许多基于模糊时间序列的预测模型,并发现论域划分机制和去模糊化预测方法是影响模型预测精度的主要因素。对此,提出了一种混合模型,将改进鲸鱼优化算法用于论域划分,提出一种基于次序和时序的平均加权算子优化去模糊化预测方法。在Alabama大学入学人数数据集上进行仿真实验,与国内外相关研究进行横向分析。结果表明,改进鲸鱼优化算法在收敛速度、求解精度都有很大程度提升;混合模型与基线模型对比时,训练阶段和预测阶段都获得了最小的均方根误差(RMSE)和平均绝对百分比误差(MAPE),训练阶段对照基线最优模型,RMSE减少21.1%,MAPE降低0.28%。
In recent years,researchers have proposed many prediction models based on fuzzy time series,and found that the domain division mechanism and the defuzzification prediction method are the main factors affecting the prediction accuracy of the model.In this paper,a hybrid prediction model is proposed.The improved whale optimization algorithm is used to divide the universe,and an average weighted operator optimization method based on order and time series is proposed to defuzzify the prediction.Through the simulation experiment on the data set of Alabama University Enrollment,the horizontal analysis is carried out with the relevant research at home and abroad.The results show that the improved whale optimization algorithm has greatly improved the convergence speed and solution accuracy;When the new model is compared with the baseline model,the minimum Root Mean Square Error(RMSE)and the Mean Absolute Percentage Error(MAPE)are obtained in the training phase and the prediction phase.Compared with the baseline optimal model,the RMSE and MAPE are reduced by 21.1% and 0.28% respectively in the training phase.
作者
汪涛
林川
郭生伟
WANG Tao;LIN Chuan;GUO Shengwei(North China Institute of Computing Technology,Beijing 100083,China;International Department,China Electronics Technology Taiji Group Corporation,Beijing 100083,China)
出处
《电子设计工程》
2023年第15期98-106,共9页
Electronic Design Engineering
关键词
鲸鱼优化算法
模糊时间序列
混沌映射
透镜成像反学习
莱维飞行
OWA
whale optimization algorithm
fuzzy time series
chaotic mapping
lens imaging reverse learning
Levy flight
Orderd Weighted Averaging
