=======================================================

KODE R: ANALISIS MULTIVARIAT (MANOVA & MANCOVA)

Dataset: Bike Sharing (Capital Bikeshare) - UCI Repo

=======================================================

1. INSTALASI DAN PEMUATAN PACKAGE

library(MVN) library(heplots) library(car) library(dplyr)

2. MEMUAT DATASET

bike_data <- read.csv(“day.csv”)

3. PRA-PEMROSESAN DAN TRANSFORMASI DATA

bike_data\(season <- as.factor(bike_data\)season) bike_data\(weathersit <- as.factor(bike_data\)weathersit)

bike_data\(log_casual <- log(bike_data\)casual + 1) bike_data\(log_registered <- log(bike_data\)registered + 1)

Y_matrix <- cbind(bike_data\(log_casual, bike_data\)log_registered)

4. STATISTIKA DESKRIPTIF

cat(“=== STATISTIKA DESKRIPTIF ===”) summary(bike_data[, c(“casual”, “registered”, “temp”)])

Menghitung standar deviasi

cat(“Deviasi Casual (Y1):”, sd(bike_data\(casual), "\n") cat("Standar Deviasi Registered (Y2): ", sd(bike_data\)registered), “”) cat(“Standar Deviasi Temp (X3):”, sd(bike_data$temp), “”)

5. PENGUJIAN ASUMSI KLASIK MULTIVARIAT

cat(“=== 1. UJI NORMALITAS MULTIVARIAT (MARDIA’S TEST) ===”) mardia_test <- mvn(data = bike_data[, c(“log_casual”, “log_registered”)], mvnTest = “mardia”) print(mardia_test$multivariateNormality)

cat(“=== 2. UJI HOMOGENITAS MATRIKS KOVARIANS (BOX’S M TEST) ===”) box_m_result <- boxM(Y_matrix ~ season * weathersit, data = bike_data) print(box_m_result) summary(box_m_result)

cat(“=== 3. UJI HOMOGENITAS KEMIRINGAN REGRESI (ASUMSI MANCOVA) ===”) model_asumsi <- manova(Y_matrix ~ temp * season * weathersit, data = bike_data) summary(model_asumsi, test = “Pillai”)

6. PEMODELAN UTAMA

cat(“=== HASIL UJI MANOVA (Tanpa Kovariat) ===”) model_manova <- manova(Y_matrix ~ season * weathersit, data = bike_data) summary(model_manova, test = “Pillai”)

cat(“=== HASIL UJI MANCOVA (Dengan Kovariat Suhu) ===”) model_mancova <- manova(Y_matrix ~ temp + season * weathersit, data = bike_data) summary(model_mancova, test = “Pillai”)

Pastikan library untuk visualisasi sudah aktif

library(MVN) library(heplots)

———————————————————

1. VISUALISASI UJI ASUMSI MULTIVARIAT

———————————————————

A. Visualisasi Normalitas Multivariat (Chi-Square Q-Q Plot)

Fungsi ini langsung menghasilkan plot visual saat uji Mardia dijalankan

hasil_mvn <- mvn(data = bike_data[, c(“log_casual”, “log_registered”)], mvnTest = “mardia”, multivariatePlot = “qq”)

B. Visualisasi Homogenitas Matriks Varians-Kovarians (Plot Box’s M)

Menampilkan dot plot sederhana dari log determinan antar kelompok uji

box_m_result <- boxM(Y_matrix ~ season * weathersit, data = bike_data) plot(box_m_result, main = “Plot Box’s M: Homogenitas Matriks Kovarians”)

C. Visualisasi Elips Sebaran (Covariance Ellipses)

Menggambar elips dari varians matriks dalam ruang variabel dependen kelompok

covEllipses(Y_matrix ~ season, data = bike_data, fill = TRUE, pooled = FALSE, main = “Elips Kovarians Berdasarkan Musim”)

———————————————————

2. VISUALISASI HASIL MANOVA & MANCOVA (HE PLOTS)

———————————————————

HE Plots (Hypothesis-Error) memproyeksikan kekuatan bukti pengujian

dengan menggunakan bentuk geometri elips untuk mewakili Hipotesis dan Error.

A. Visualisasi MANOVA

heplot(model_manova, fill = TRUE, fill.alpha = 0.1, main = “HE Plot MANOVA: Efek Musim & Cuaca”, xlab = “Log Pengguna Kasual (Y1)”, ylab = “Log Pengguna Terdaftar (Y2)”)

B. Visualisasi MANCOVA (Dengan Kovariat ‘temp’)

heplot(model_mancova, fill = TRUE, fill.alpha = 0.1, main = “HE Plot MANCOVA: Intervensi Suhu Lingkungan”, xlab = “Log Pengguna Kasual (Y1)”, ylab = “Log Pengguna Terdaftar (Y2)”)