基于多模型的时间序列需求预测与模拟分析  

Time Series Demand Forecasting and Simulation Analysis Based on Multi-Model Approach

作　　者：刘新芬 LIU Xinfen(Liaocheng High-Tech Business Incubation Service Center,Liaocheng Shandong 252000,China)

机构地区：[1]聊城市高新技术创业服务中心,山东聊城252000

出　　处：《信息与电脑》2025年第2期122-124,共3页Information & Computer

摘　　要：文章构建一套考虑了多模型的分析框架,用于时间序列数据中的需求预测问题,同时结合蒙特卡洛模拟方法对预测结果进行预测稳定性分析。研究采用多种模型并结合特征工程,对包含多变量的模拟数据进行建模分析。实验结果显示,岭回归模型在预测性能上综合表现最佳。文章方法有效结合了特征工程与模型优化,为时间序列需求预测提供了新的解决方案。The study aims to address the demand forecasting problem in time series data by constructing an ensemble analysis framework considering multi-models.Additionally,Monte Carlo simulation is incorporated to analyze the predictive stability of the results.The research integrates various models with feature engineering techniques to model multivariate simulated data.Experimental results indicate that the ridge regression model demonstrates the best overall predictive performance.This approach effectively integrates feature engineering and model optimization,providing a novel solution for time series demand forecasting.

关 键 词：机器学习 需求预测 财务应用 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]