Multiple linear regression model (AICc Based)

Pre Requirements:- You will need data in a CSV as follows: 

# Install the MuMIn package (for model selection using AIC/AICc)

install.packages("MuMIn")  # Only run this once

library(MuMIn)             # Load the MuMIn package

# Load your dataset from a CSV file

mydata <- read.csv("C:/Users/Rounak Choudhary/Desktop/BirdAIC.csv")

# Step 1: Define a global linear model

# Fspe is the response variable (e.g., bird species richness)

# The model includes multiple landscape/environmental predictors

global.model <- lm(Fspe ~ Area + AIV + Tree + Shrub + Grass + Crop + Built + Bare +

                   Water + Cont + PD + ED + IJI + SDI + CA, data = mydata)

# Step 2: Ensure the model selection function handles missing values properly

options(na.action = "na.fail")  # Required for dredge() to function correctly

# Step 3: Use dredge() to generate all possible models based on subsets of predictors

model.set <- dredge(global.model)

# Step 4: Print the full model selection table

# This includes models ranked by AICc, along with their variables, AICc scores, weights, etc.

print(model.set)

# Step 5: Extract the best model (i.e., the model with the lowest AICc)

best.model <- get.models(model.set, 1)[[1]]  # Extract top-ranked model

# Display a detailed summary of the best model (coefficients, p-values, R², etc.)

summary(best.model)

# Compare the AIC of the global model (for context)

AIC(global.model)

# Show the top-ranked model from the selection table (can change 1:5 to see more top models)

model.set[1:1, ]

# Step 6: Extract the top 3 models from the model set based on AICc

top3.models <- get.models(model.set, subset = 1:3)

# Step 7: Display a summary of each of the top 3 models

# lapply applies summary() to each model in the list

lapply(top3.models, summary)