Segmented Analysis

Pre Requirements:- The data should be prepared as follows: 

# Install the required packages (only needs to be done once)

install.packages("segmented")  # For segmented (piecewise) regression modeling

install.packages("ggplot2")    # For advanced plotting (not directly used in this script but useful)

# Load the installed packages into the R session

library(segmented)  # Loads functions for segmented regression

library(ggplot2)    # Loads functions for advanced graphics (not used in this script)

# Read a CSV file into a data frame. File should have at least two columns: ShrubP and Richness

df <- read.csv("C:/Users/Rounak Choudhary/Desktop/example.csv")

# Fit a basic linear regression model with Richness as the response and ShrubP as the predictor

lm_model <- lm(Richness ~ ShrubP, data = df)

# Display the minimum and maximum values of the ShrubP variable to help choose a reasonable breakpoint guess

range(df$ShrubP)

# Fit a segmented (piecewise linear) regression model with a breakpoint in ShrubP

# seg.Z = ~ ShrubP tells the function to search for a breakpoint in ShrubP

# psi = 27 gives an initial guess for the breakpoint (must be inside the range of ShrubP)

seg_model <- segmented(lm_model, seg.Z = ~ ShrubP, psi = 27)

# Print a detailed summary of the segmented model including:

# - Regression coefficients

# - Estimated breakpoint (psi)

# - Confidence intervals and significance tests

summary(seg_model)

# Extract the estimated breakpoint value from the model

breakpoint <- seg_model$psi[2]  # This is the estimated value of ShrubP where the slope changes

# Create a base scatter plot of Richness vs ShrubP

plot(df$ShrubP, df$Richness, pch = 16, main = "Threshold in Tree Cover")

# Add the fitted segmented regression line to the existing plot in red

plot(seg_model, add = TRUE, col = "red")

# Add a vertical dashed blue line at the estimated breakpoint location

abline(v = breakpoint, col = "blue", lty = 2)