Perbandingan Metode Prediksi untuk Nilai Jual USD: Holt-Winters, Holt's, dan Single Exponential Smoothing

Author:

Rosita Yesy Diah,Moonlight Lady Silk

Abstract

In the ever-changing landscape of the global economy, the role of the United States Dollar (USD) as the backbone of the international financial system significantly influences market stability and dynamics. The close correlation between fluctuations in the USD exchange rate and internal and external factors demands effective prediction methods to understand and manage associated risks. This study aims to compare the performance of three main prediction methods: Single Exponential Smoothing (SES), Holt's Method, and Holt-Winters Method, in forecasting USD exchange rates. Utilizing historical data from the Central Statistics Agency (BPS) and testing under three training data distribution scenarios (45%, 55%, and 75%), this research provides in-depth findings on the strengths and weaknesses of each prediction method. Performance evaluations include the time required, Mean Absolute Error (MAE), Mean Squared Error (MSE), Mean Absolute Percentage Error (MAPE), R-Squared, and correlation for the implementation of each method. If averaged, the results are as follows for SES, Holt’s, and Holt’s Winter, respectively: SES (1.58; 284.20; 68,768.26; 440.07; 0.03; -2.12; Nan), Holt’s (1.39; 890.23; 426,377.44; 1,043.28; 0.06; -24.28; -0.66), and Holt’s Winter (1.20; 997.45; 513,657.58; 1,168.00; 0.07; -30.62; -1.55). Overall, this indicates that the Holt-Winters Method stands out with significant performance, especially in scenarios with larger training data distributions, with a low R-Squared value (-4.618) and satisfactory correlation (0.417). Holt's Method also shows improved accuracy, while Single Exponential Smoothing (SES) offers time efficiency, albeit with limitations in explaining data variations. In conclusion, this research provides valuable guidance for business stakeholders, investors, and policymakers in selecting prediction methods suitable for their data characteristics and analysis goals, with the potential for a positive impact on business strategies, competitiveness, and risk management amid the uncertainty of USD exchange rate fluctuations.

Publisher

Sekawan Institute

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3