粒度
计算机科学
超参数
比例(比率)
数据挖掘
平衡(能力)
系列(地层学)
机器学习
人工智能
数据科学
地图学
地理
程序设计语言
医学
古生物学
物理医学与康复
生物
作者
Maxime C. Cohen,Paul-Emile Gras,Arthur Pentecoste,Renyu Zhang
出处
期刊:Springer series in supply chain management
日期:2022-01-01
卷期号:: 129-149
标识
DOI:10.1007/978-3-030-85855-1_7
摘要
In this chapter, we discuss two more advanced methods, which are recent advancements in demand prediction. We first present the Prophet method, which is an open-sourced library released by Facebook researchers in 2017. Prophet is a time-series demand prediction method that often performs well on large-scale problems. We explain the method and discuss its implementation both with and without incorporating features. We then present a method that can strike a good balance between data aggregation (i.e., finding the right data granularity level) and demand prediction accuracy. We present the method, discuss how to fine-tune its hyperparameters, and conclude by interpreting the results obtained on the accompanying dataset.
科研通智能强力驱动
Strongly Powered by AbleSci AI