grey_models: Grey Prediction Models

In zhjx19/mathmodels: Comprehensive Mathematical Modeling Algorithms in R

Grey Prediction Models

Description

Implements grey prediction models for time series forecasting: GM11 applies the GM(1,1) model with level ratio test. GM1N applies the GM(1,N) model with multiple related factors. DGM21 applies the DGM(2,1) model for second-order dynamics. verhulst applies the Verhulst model for logistic growth.

Usage

GM11(X)

<<<<<<< HEAD
GM1N(dat, new_data = NULL)
=======
GM1N(dat)
>>>>>>> 3e92d3d55418301db33a496b2db922076ea97b15

DGM21(X)

verhulst(X)

Arguments

X	For GM11, DGM21, verhulst: Numeric vector of original time series data.
dat	For GM1N: Data frame or matrix, last column is characteristic series, others are related factors.

Value

For GM11: List with fitted values (fitted), next prediction (pnext), prediction function (f), matrix (mat), parameters (u), level ratios (lambda), and range (rng). For GM1N: List with fitted values (fitted), posterior variance ratio (C), small error probability (P), and prediction function (f). For DGM21, verhulst: List with fitted values (fitted), next prediction (pnext), prediction function (f), matrix (mat), and parameters (u).

Examples

# Sample time series for GM11, DGM21, Verhulst
x = c(100, 120, 145, 175, 210)

# GM11
result = GM11(x)
result$fitted    # Fitted values
result$pnext     # Next prediction
result$f(6:8)    # Predict next 3 periods

# DGM21
x = c(2.874,3.278,3.39,3.679,3.77,3.8)
result = DGM21(x)
result$fitted    # Fitted values
result$pnext     # Next prediction
result$f(6:8)    # Predict next 3 periods

# Verhulst
x = c(4.93,2.33,3.87,4.35,6.63,7.15,5.37,6.39,7.81,8.35)
result = verhulst(x)
result$fitted    # Fitted values
result$pnext     # Next prediction
result$f(6:8)    # Predict next 3 periods

# Sample data for GM1N
data = data.frame(
  factor1 = c(50, 55, 60, 65, 70),
  factor2 = c(20, 22, 25, 28, 30),
  output = c(100, 120, 145, 175, 210)
)
result = GM1N(data)
result$fitted

zhjx19/mathmodels documentation built on June 2, 2025, 12:18 a.m.