非线性回归
线性回归
多项式回归
置信度和预测带
置信区间
回归分析
统计
回归
数据集
非线性系统
数学
多项式的
应用数学
线性模型
数学分析
物理
量子力学
作者
C. Prieto,Leonardo Guzmán Alegría
标识
DOI:10.1080/00032719.2023.2278662
摘要
Confidence bands are used to estimate the likely range of the true parameter, whereas prediction bands provide an idea of the possible values of the parameter after applying a regression model. This study proposes an equation to construct confidence and prediction bands applicable to any linear or nonlinear chemometric regression model; that is, statistical methods used in analytical chemistry to establish relationships between two or more variables. The proposed equation is developed and presented as a useful tool to represent the uncertainty in estimates based on a set of data in chemical measurements. Once the set of data is adjusted as a regression model (linear, polynomial, or multivariable), the proposed equation can be used to construct confidence and prediction bands based on the adjustment equation. The proposed equation operates if the adjustment equation is continuous and derivable. The plot bands on any regression model provide a solid understanding of uncertainty assessment and have the advantage of total control and independence of algorithms and numeral approximations.
科研通智能强力驱动
Strongly Powered by AbleSci AI