Tugas Minggu 6
Input Data
## Warning: package 'readxl' was built under R version 4.3.2## Warning: package 'tidyverse' was built under R version 4.3.2## Warning: package 'ggplot2' was built under R version 4.3.2## Warning: package 'tibble' was built under R version 4.3.2## Warning: package 'tidyr' was built under R version 4.3.2## Warning: package 'readr' was built under R version 4.3.2## Warning: package 'purrr' was built under R version 4.3.2## Warning: package 'dplyr' was built under R version 4.3.2## Warning: package 'forcats' was built under R version 4.3.2## Warning: package 'lubridate' was built under R version 4.3.2## ── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
## ✔ dplyr 1.1.3 ✔ readr 2.1.4
## ✔ forcats 1.0.0 ✔ stringr 1.5.1
## ✔ ggplot2 3.4.4 ✔ tibble 3.2.1
## ✔ lubridate 1.9.3 ✔ tidyr 1.3.0
## ✔ purrr 1.0.2
## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::filter() masks stats::filter()
## ✖ dplyr::lag() masks stats::lag()
## ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errorsdata <- read_excel("C:/Users/Admin/Downloads/PSD Kelompok 3 (3).xlsx")
names(data)[names(data) == "Jumlah Ulasan"] <- "J.Ulasan"
head(data)## # A tibble: 6 × 6
## Brand HARGA LOKASI J.Ulasan RAM Penyimpanan
## <chr> <dbl> <chr> <chr> <chr> <chr>
## 1 iphone 4388000 Jabodetabek 10 4 256
## 2 iphone 9458000 Jabodetabek 3071 4 128
## 3 iphone 9409000 Jabodetabek 2713 4 128
## 4 iphone 5047000 Luar Jawa 87 4 64
## 5 iphone 11287000 Jabodetabek 567 6 128
## 6 iphone 21447000 Jabodetabek 732 8 256Pada tahap ini, Kami akan mengambil data hasil scrapping pada Minggu 3. Data ini adalah Data Penjualan Handphone, dengan Harga sebagai peubah Respon dan 5 Peubah Penjelas. Peubah Penjelas terdiri dari Brand, Lokasi, Ram, Penyimpanan.
Pada Tahap ini Toko tidak kami masukkan ke model atas dasar pertimbangan pengelompokkan yang rumit untuk dilakukan sehingga pada bagian ini kami tidak memasukkan peubah Toko terlebih dahulu dan memodelkan dengan 5 Peubah Penjelas
data$Brand <- relevel(as.factor(data$Brand), ref="iphone")
data$LOKASI <- relevel(as.factor(data$LOKASI), ref="Jabodetabek")
data$RAM <- relevel(as.factor(as.numeric(data$RAM)), ref = "6")
data$Penyimpanan <- relevel(as.factor(as.numeric(data$Penyimpanan)), ref = "128")
data$J.Ulasan <- as.numeric(data$J.Ulasan)Pada tahapan ini , kami melakukan Re-level pada data-data kategorik yaitu Brand, Lokasi , RAM dan Penyimpanan. RAM dan Penyimpanan memang berbentuk data numerik, namun dalam substansinya, angka-angka pada RAM dan Penyimpanan sudah disetting dalam beberapa kategori sehingga dapat dikatakan bahwa data ini lebih ke Kategorik daripada Numerik.
Pada Tahap ini, Kami juga menentukan Reference yaitu Untuk Brand Kami menerapkan Reference di kategori Iphone, lalu untuk Lokasi kamu menerapkan reference di Jabodetabek, 6 Untuk Ram dan 128 untuk Penyimpanan.
Terakhir kami menyiapkan data ulasan sebagai numeric. Hal ini dilakukan agar data tetap dibaca sebagai numerik pada model.
Visualisasi Dasar
Untuk Visualisasi Sebelum Ke Model, Dapat dilihat bahwa Kecenderungan Harga menjulur ke kanan. Hal ini mengartikan bahwa harga Handphone menyebar di daerah 5 Jutaan. Kemudian data mulai turun secara signifikan sampai ke harga puluhan juta. Sebaran data hasil histogram ini juga mengindikasikan adanya outlier karena data tidak terdistribusi secara merata, untuk itu kami akan melakukan pendeteksian pencilan setelah pemodelan.
##
## Attaching package: 'MASS'## The following object is masked from 'package:dplyr':
##
## select##
## Call:
## lm(formula = HARGA ~ Brand + LOKASI + RAM + Penyimpanan + J.Ulasan,
## data = data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -12853595 -1607087 -586536 1435603 14613759
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 1.102e+07 1.678e+05 65.685 < 2e-16 ***
## Brandasus -7.748e+06 4.688e+05 -16.529 < 2e-16 ***
## Brandinfinix -9.997e+06 2.215e+05 -45.142 < 2e-16 ***
## Brandoppo -8.896e+06 2.553e+05 -34.848 < 2e-16 ***
## BrandRealme -8.830e+06 2.508e+05 -35.211 < 2e-16 ***
## BrandSamsung -7.186e+06 1.737e+05 -41.364 < 2e-16 ***
## Brandvivo -8.731e+06 1.980e+05 -44.103 < 2e-16 ***
## Brandxiaomi -9.110e+06 1.778e+05 -51.228 < 2e-16 ***
## LOKASIBali 3.884e+05 4.681e+05 0.830 0.406761
## LOKASIDI Yogyakarta -6.591e+05 7.483e+05 -0.881 0.378479
## LOKASIJawa Barat -1.784e+04 2.283e+05 -0.078 0.937735
## LOKASIJawa Tengah -1.905e+05 4.780e+05 -0.399 0.690263
## LOKASIJawa Timur 3.799e+05 1.773e+05 2.142 0.032244 *
## LOKASILuar Jawa 4.378e+05 1.671e+05 2.620 0.008841 **
## RAM1 -1.117e+07 1.327e+06 -8.413 < 2e-16 ***
## RAM2 -8.142e+06 5.753e+05 -14.153 < 2e-16 ***
## RAM3 -6.381e+06 3.622e+05 -17.618 < 2e-16 ***
## RAM4 -2.780e+06 1.822e+05 -15.259 < 2e-16 ***
## RAM8 9.669e+05 1.661e+05 5.820 6.50e-09 ***
## RAM12 4.762e+06 2.669e+05 17.842 < 2e-16 ***
## RAM16 8.399e+06 6.345e+05 13.237 < 2e-16 ***
## RAM18 1.271e+07 2.100e+06 6.053 1.59e-09 ***
## RAM24 1.567e+06 1.742e+06 0.899 0.368505
## Penyimpanan8 -3.033e+06 2.928e+06 -1.036 0.300349
## Penyimpanan16 2.474e+06 1.156e+06 2.140 0.032437 *
## Penyimpanan32 3.742e+06 6.041e+05 6.194 6.63e-10 ***
## Penyimpanan64 -5.358e+05 1.917e+05 -2.795 0.005226 **
## Penyimpanan256 5.122e+05 1.379e+05 3.714 0.000207 ***
## Penyimpanan512 2.456e+06 2.494e+05 9.848 < 2e-16 ***
## J.Ulasan -3.379e+02 2.808e+02 -1.203 0.228918
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 2870000 on 3091 degrees of freedom
## Multiple R-squared: 0.6273, Adjusted R-squared: 0.6239
## F-statistic: 179.4 on 29 and 3091 DF, p-value: < 2.2e-16## Analysis of Variance Table
##
## Response: HARGA
## Df Sum Sq Mean Sq F value Pr(>F)
## Brand 7 2.2663e+16 3.2376e+15 392.9550 < 2e-16 ***
## LOKASI 6 1.2717e+14 2.1196e+13 2.5726 0.01732 *
## RAM 9 1.8693e+16 2.0770e+15 252.0915 < 2e-16 ***
## Penyimpanan 6 1.3774e+15 2.2957e+14 27.8633 < 2e-16 ***
## J.Ulasan 1 1.1932e+13 1.1932e+13 1.4481 0.22892
## Residuals 3091 2.5467e+16 8.2392e+12
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Didapatkan Model Sebagai Berikut ini , Dimana R-Squared 62%. Didapat pula bahwa mayoritas peubah yang digunaka signfikan dalam taraf kepercayaan 95 %. Namun ada beberapa peubah juga yang tidak signifkan seperti Jumlah Ulasan.
## Warning in sqrt(crit * p * (1 - hh)/hh): NaNs produced
## Warning in sqrt(crit * p * (1 - hh)/hh): NaNs produced
## Warning: package 'lmtest' was built under R version 4.3.3## Loading required package: zoo## Warning: package 'zoo' was built under R version 4.3.2##
## Attaching package: 'zoo'## The following objects are masked from 'package:base':
##
## as.Date, as.Date.numeric##
## Durbin-Watson test
##
## data: model
## DW = 1.9868, p-value = 0.3095
## alternative hypothesis: true autocorrelation is greater than 0Dari Plot Residual Vs Fitted, Asumsi Homoskedastisitas tidak terpenuhi karena banyaknya residual data hasil tidak menyebar di nilai tengah dimana y= 0 . Sehingga dapat disimpulkan bahwa terjadi Heteroskedasitas yang artinya ragam sisanya tidak homogen.
Dari Plot QQ-Residual , walaupun beberapa data menyebar normal di tengah, namun, mayoritas data di ujung dan di belakannya tidak berada di garis kenormalan data. Sehinga dapat dikatakan data tidak menyebar normal namun harus tetap dilakukan uji secara formal. Data ini bisa berpotensi tidak normal karena adanya pencilan. Data tidak menyebar mengikuti garis lurus. Bagian atas yang melenceng dari garis menunjukkan bahwa sisaan model menjulur ke kanan.
Dari Plot Fitted Values Vs Standarad Residual dapat dilihat bahwa data cenderung menyebar dan membentuk pola, hal ini mengindikasikan bahwa nilai harapan sisaan pada data tidak saling bebas . karena adanya pola yang terbentuk.
## Warning: package 'car' was built under R version 4.3.2## Loading required package: carData## Warning: package 'carData' was built under R version 4.3.2##
## Attaching package: 'car'## The following object is masked from 'package:dplyr':
##
## recode## The following object is masked from 'package:purrr':
##
## some## GVIF Df GVIF^(1/(2*Df))
## Brand 2.494361 7 1.067466
## LOKASI 1.130748 6 1.010293
## RAM 7.462211 9 1.118131
## Penyimpanan 4.322606 6 1.129741
## J.Ulasan 1.067401 1 1.033151## Warning: package 'leaps' was built under R version 4.3.3VIF pada Seluruh Data Berada dibawah 10, yang artinya cukup bukti untuk mengatakan tidak ada multikolinearitas.
Pemilihan Peubah
Best Subset.
regfit.full <- regsubsets(HARGA ~ Brand+ LOKASI + RAM + Penyimpanan+ J.Ulasan, data=data , nvmax = 29)
reg.summary <- summary(regfit.full)
summary(regfit.full)## Subset selection object
## Call: regsubsets.formula(HARGA ~ Brand + LOKASI + RAM + Penyimpanan +
## J.Ulasan, data = data, nvmax = 29)
## 29 Variables (and intercept)
## Forced in Forced out
## Brandasus FALSE FALSE
## Brandinfinix FALSE FALSE
## Brandoppo FALSE FALSE
## BrandRealme FALSE FALSE
## BrandSamsung FALSE FALSE
## Brandvivo FALSE FALSE
## Brandxiaomi FALSE FALSE
## LOKASIBali FALSE FALSE
## LOKASIDI Yogyakarta FALSE FALSE
## LOKASIJawa Barat FALSE FALSE
## LOKASIJawa Tengah FALSE FALSE
## LOKASIJawa Timur FALSE FALSE
## LOKASILuar Jawa FALSE FALSE
## RAM1 FALSE FALSE
## RAM2 FALSE FALSE
## RAM3 FALSE FALSE
## RAM4 FALSE FALSE
## RAM8 FALSE FALSE
## RAM12 FALSE FALSE
## RAM16 FALSE FALSE
## RAM18 FALSE FALSE
## RAM24 FALSE FALSE
## Penyimpanan8 FALSE FALSE
## Penyimpanan16 FALSE FALSE
## Penyimpanan32 FALSE FALSE
## Penyimpanan64 FALSE FALSE
## Penyimpanan256 FALSE FALSE
## Penyimpanan512 FALSE FALSE
## J.Ulasan FALSE FALSE
## 1 subsets of each size up to 29
## Selection Algorithm: exhaustive
## Brandasus Brandinfinix Brandoppo BrandRealme BrandSamsung Brandvivo
## 1 ( 1 ) " " " " " " " " " " " "
## 2 ( 1 ) " " " " " " " " " " " "
## 3 ( 1 ) " " "*" " " " " " " " "
## 4 ( 1 ) " " "*" " " " " " " "*"
## 5 ( 1 ) " " "*" " " "*" " " "*"
## 6 ( 1 ) " " "*" "*" "*" "*" "*"
## 7 ( 1 ) " " "*" "*" "*" "*" "*"
## 8 ( 1 ) " " "*" "*" "*" "*" "*"
## 9 ( 1 ) " " "*" "*" "*" "*" "*"
## 10 ( 1 ) " " "*" "*" "*" "*" "*"
## 11 ( 1 ) " " "*" "*" "*" "*" "*"
## 12 ( 1 ) "*" "*" "*" "*" "*" "*"
## 13 ( 1 ) "*" "*" "*" "*" "*" "*"
## 14 ( 1 ) "*" "*" "*" "*" "*" "*"
## 15 ( 1 ) "*" "*" "*" "*" "*" "*"
## 16 ( 1 ) "*" "*" "*" "*" "*" "*"
## 17 ( 1 ) "*" "*" "*" "*" "*" "*"
## 18 ( 1 ) "*" "*" "*" "*" "*" "*"
## 19 ( 1 ) "*" "*" "*" "*" "*" "*"
## 20 ( 1 ) "*" "*" "*" "*" "*" "*"
## 21 ( 1 ) "*" "*" "*" "*" "*" "*"
## 22 ( 1 ) "*" "*" "*" "*" "*" "*"
## 23 ( 1 ) "*" "*" "*" "*" "*" "*"
## 24 ( 1 ) "*" "*" "*" "*" "*" "*"
## 25 ( 1 ) "*" "*" "*" "*" "*" "*"
## 26 ( 1 ) "*" "*" "*" "*" "*" "*"
## 27 ( 1 ) "*" "*" "*" "*" "*" "*"
## 28 ( 1 ) "*" "*" "*" "*" "*" "*"
## 29 ( 1 ) "*" "*" "*" "*" "*" "*"
## Brandxiaomi LOKASIBali LOKASIDI Yogyakarta LOKASIJawa Barat
## 1 ( 1 ) " " " " " " " "
## 2 ( 1 ) "*" " " " " " "
## 3 ( 1 ) "*" " " " " " "
## 4 ( 1 ) "*" " " " " " "
## 5 ( 1 ) "*" " " " " " "
## 6 ( 1 ) "*" " " " " " "
## 7 ( 1 ) "*" " " " " " "
## 8 ( 1 ) "*" " " " " " "
## 9 ( 1 ) "*" " " " " " "
## 10 ( 1 ) "*" " " " " " "
## 11 ( 1 ) "*" " " " " " "
## 12 ( 1 ) "*" " " " " " "
## 13 ( 1 ) "*" " " " " " "
## 14 ( 1 ) "*" " " " " " "
## 15 ( 1 ) "*" " " " " " "
## 16 ( 1 ) "*" " " " " " "
## 17 ( 1 ) "*" " " " " " "
## 18 ( 1 ) "*" " " " " " "
## 19 ( 1 ) "*" " " " " " "
## 20 ( 1 ) "*" " " " " " "
## 21 ( 1 ) "*" " " " " " "
## 22 ( 1 ) "*" " " " " " "
## 23 ( 1 ) "*" " " " " " "
## 24 ( 1 ) "*" " " " " " "
## 25 ( 1 ) "*" " " " " " "
## 26 ( 1 ) "*" " " "*" " "
## 27 ( 1 ) "*" "*" "*" " "
## 28 ( 1 ) "*" "*" "*" " "
## 29 ( 1 ) "*" "*" "*" "*"
## LOKASIJawa Tengah LOKASIJawa Timur LOKASILuar Jawa RAM1 RAM2 RAM3
## 1 ( 1 ) " " " " " " " " " " " "
## 2 ( 1 ) " " " " " " " " " " " "
## 3 ( 1 ) " " " " " " " " " " " "
## 4 ( 1 ) " " " " " " " " " " " "
## 5 ( 1 ) " " " " " " " " " " " "
## 6 ( 1 ) " " " " " " " " " " " "
## 7 ( 1 ) " " " " " " " " " " " "
## 8 ( 1 ) " " " " " " " " " " " "
## 9 ( 1 ) " " " " " " " " " " "*"
## 10 ( 1 ) " " " " " " " " "*" "*"
## 11 ( 1 ) " " " " " " " " "*" "*"
## 12 ( 1 ) " " " " " " " " "*" "*"
## 13 ( 1 ) " " " " " " " " "*" "*"
## 14 ( 1 ) " " " " " " "*" "*" "*"
## 15 ( 1 ) " " " " " " "*" "*" "*"
## 16 ( 1 ) " " " " " " "*" "*" "*"
## 17 ( 1 ) " " " " " " "*" "*" "*"
## 18 ( 1 ) " " " " " " "*" "*" "*"
## 19 ( 1 ) " " " " " " "*" "*" "*"
## 20 ( 1 ) " " " " "*" "*" "*" "*"
## 21 ( 1 ) " " "*" "*" "*" "*" "*"
## 22 ( 1 ) " " "*" "*" "*" "*" "*"
## 23 ( 1 ) " " "*" "*" "*" "*" "*"
## 24 ( 1 ) " " "*" "*" "*" "*" "*"
## 25 ( 1 ) " " "*" "*" "*" "*" "*"
## 26 ( 1 ) " " "*" "*" "*" "*" "*"
## 27 ( 1 ) " " "*" "*" "*" "*" "*"
## 28 ( 1 ) "*" "*" "*" "*" "*" "*"
## 29 ( 1 ) "*" "*" "*" "*" "*" "*"
## RAM4 RAM8 RAM12 RAM16 RAM18 RAM24 Penyimpanan8 Penyimpanan16
## 1 ( 1 ) " " " " " " " " " " " " " " " "
## 2 ( 1 ) " " " " " " " " " " " " " " " "
## 3 ( 1 ) " " " " " " " " " " " " " " " "
## 4 ( 1 ) " " " " " " " " " " " " " " " "
## 5 ( 1 ) " " " " " " " " " " " " " " " "
## 6 ( 1 ) " " " " " " " " " " " " " " " "
## 7 ( 1 ) " " " " " " " " " " " " " " " "
## 8 ( 1 ) " " " " "*" " " " " " " " " " "
## 9 ( 1 ) "*" " " "*" " " " " " " " " " "
## 10 ( 1 ) "*" " " "*" " " " " " " " " " "
## 11 ( 1 ) "*" " " "*" " " " " " " " " " "
## 12 ( 1 ) "*" " " "*" "*" " " " " " " " "
## 13 ( 1 ) "*" " " "*" "*" " " " " " " " "
## 14 ( 1 ) "*" " " "*" "*" " " " " " " " "
## 15 ( 1 ) "*" "*" "*" "*" " " " " " " " "
## 16 ( 1 ) "*" "*" "*" "*" " " " " " " " "
## 17 ( 1 ) "*" "*" "*" "*" "*" " " " " " "
## 18 ( 1 ) "*" "*" "*" "*" "*" " " " " " "
## 19 ( 1 ) "*" "*" "*" "*" "*" " " " " " "
## 20 ( 1 ) "*" "*" "*" "*" "*" " " " " " "
## 21 ( 1 ) "*" "*" "*" "*" "*" " " " " " "
## 22 ( 1 ) "*" "*" "*" "*" "*" " " " " "*"
## 23 ( 1 ) "*" "*" "*" "*" "*" " " " " "*"
## 24 ( 1 ) "*" "*" "*" "*" "*" " " "*" "*"
## 25 ( 1 ) "*" "*" "*" "*" "*" "*" "*" "*"
## 26 ( 1 ) "*" "*" "*" "*" "*" "*" "*" "*"
## 27 ( 1 ) "*" "*" "*" "*" "*" "*" "*" "*"
## 28 ( 1 ) "*" "*" "*" "*" "*" "*" "*" "*"
## 29 ( 1 ) "*" "*" "*" "*" "*" "*" "*" "*"
## Penyimpanan32 Penyimpanan64 Penyimpanan256 Penyimpanan512 J.Ulasan
## 1 ( 1 ) " " " " " " "*" " "
## 2 ( 1 ) " " " " " " "*" " "
## 3 ( 1 ) " " " " " " "*" " "
## 4 ( 1 ) " " " " " " "*" " "
## 5 ( 1 ) " " " " " " "*" " "
## 6 ( 1 ) " " " " " " " " " "
## 7 ( 1 ) " " "*" " " " " " "
## 8 ( 1 ) " " "*" " " " " " "
## 9 ( 1 ) " " " " " " " " " "
## 10 ( 1 ) " " " " " " " " " "
## 11 ( 1 ) " " " " " " "*" " "
## 12 ( 1 ) " " " " " " " " " "
## 13 ( 1 ) " " " " " " "*" " "
## 14 ( 1 ) " " " " " " "*" " "
## 15 ( 1 ) " " " " " " "*" " "
## 16 ( 1 ) "*" " " " " "*" " "
## 17 ( 1 ) "*" " " " " "*" " "
## 18 ( 1 ) "*" " " "*" "*" " "
## 19 ( 1 ) "*" "*" "*" "*" " "
## 20 ( 1 ) "*" "*" "*" "*" " "
## 21 ( 1 ) "*" "*" "*" "*" " "
## 22 ( 1 ) "*" "*" "*" "*" " "
## 23 ( 1 ) "*" "*" "*" "*" "*"
## 24 ( 1 ) "*" "*" "*" "*" "*"
## 25 ( 1 ) "*" "*" "*" "*" "*"
## 26 ( 1 ) "*" "*" "*" "*" "*"
## 27 ( 1 ) "*" "*" "*" "*" "*"
## 28 ( 1 ) "*" "*" "*" "*" "*"
## 29 ( 1 ) "*" "*" "*" "*" "*"Pada Subsetting kali ini , kami memilih nilai nvmax 29 untuk memasukkan semua variabel, meningat sebagian besar variabel kami adalah variabel dummy. Setelah dilihat dari summaryna, ternyata model ke 29 atau model maksimal adalah model yang memasukkan semua variabel dummy tanpa adanya variabel dummy yang terbuang, hal ini menjadi pertimbangan kami untuk memilih model ini, namun perlu dilakukan pengetesan.
## [1] 0.09240237 0.13856492 0.17996271 0.22998662 0.27180073 0.33017864
## [7] 0.40556025 0.46406155 0.51016061 0.53845534 0.56307456 0.58207056
## [13] 0.59511227 0.60330756 0.61001694 0.61525926 0.61958078 0.62152828
## [19] 0.62252194 0.62318433 0.62365766 0.62410094 0.62415064 0.62415792
## [25] 0.62413529 0.62410899 0.62407385 0.62397130 0.62385039## [1] 24Dari adj R^2 dapat dilihat bahwa model ke 24 mendapat r-squared tertinggi di 0.624 namun perlu diperhatikan bahwa model ini tidak menggapai seluruh variabel dummy.
## (Intercept) Brandasus Brandinfinix Brandoppo
## 1.102538e+07 -7.643575e+06 -1.000860e+07 -8.898051e+06
## BrandRealme BrandSamsung Brandvivo Brandxiaomi
## -8.830088e+06 -7.192492e+06 -8.728234e+06 -9.105534e+06
## LOKASIJawa Timur LOKASILuar Jawa RAM1 RAM2
## 3.830273e+05 4.408839e+05 -1.117701e+07 -8.169859e+06
## RAM3 RAM4 RAM8 RAM12
## -6.385813e+06 -2.787857e+06 9.568391e+05 4.737387e+06
## RAM16 RAM18 Penyimpanan8 Penyimpanan16
## 8.300965e+06 1.256514e+07 -3.012885e+06 2.487205e+06
## Penyimpanan32 Penyimpanan64 Penyimpanan256 Penyimpanan512
## 3.748907e+06 -5.397252e+05 5.161355e+05 2.492005e+06
## J.Ulasan
## -3.327357e+02## [1] 4408.72093 4025.65492 3682.31799 3267.74290 2921.42153 2438.19702
## [7] 1814.56096 1330.98154 950.28778 717.04507 514.33213 358.21288
## [13] 251.37716 184.62932 130.19089 87.89424 53.22162 38.14977
## [19] 30.95178 26.48923 23.58780 20.93643 21.52792 22.46879
## [25] 23.65580 24.87283 26.16250 28.00610 30.00000## [1] 22Dari Nilai Cp-Mallows, didapat 20.936 di model ke 22 yang artinya paling rendah, namun sama seperti kasus sebelumnya, model ini tidak mencakup seluruh variabel dummy.
## [1] 95587.32 95423.40 95268.69 95071.25 94895.99 94634.19 94260.56 93936.23
## [9] 93654.51 93467.82 93295.73 93156.00 93056.05 92991.22 92936.98 92893.74
## [17] 92857.48 92840.45 92831.24 92824.75 92819.82 92815.14 92813.72 92812.65
## [25] 92811.83 92811.04 92810.32 92810.16 92810.16## [1] 29AIC terendah di 29, dan di model 29 semua model terambil, sehingga ini bisa menjadi pertimbangan dalam pemilihan subsetnya.
## (Intercept) Brandasus Brandinfinix Brandoppo
## 1.102039e+07 -7.748241e+06 -9.996847e+06 -8.896092e+06
## BrandRealme BrandSamsung Brandvivo Brandxiaomi
## -8.829891e+06 -7.185516e+06 -8.730847e+06 -9.109551e+06
## LOKASIBali LOKASIDI Yogyakarta LOKASIJawa Barat LOKASIJawa Tengah
## 3.883858e+05 -6.590816e+05 -1.783686e+04 -1.904907e+05
## LOKASIJawa Timur LOKASILuar Jawa RAM1 RAM2
## 3.798504e+05 4.377656e+05 -1.116853e+07 -8.141921e+06
## RAM3 RAM4 RAM8 RAM12
## -6.381449e+06 -2.780065e+06 9.668636e+05 4.762440e+06
## RAM16 RAM18 RAM24 Penyimpanan8
## 8.398689e+06 1.271118e+07 1.567127e+06 -3.032647e+06
## Penyimpanan16 Penyimpanan32 Penyimpanan64 Penyimpanan256
## 2.473774e+06 3.741696e+06 -5.357751e+05 5.121893e+05
## Penyimpanan512 J.Ulasan
## 2.456009e+06 -3.379496e+02## [1] -287.5025 -443.3781 -590.0422 -779.4395 -946.6506 -1200.4091
## [7] -1565.9870 -1882.2796 -2155.9450 -2334.5980 -2498.6373 -2630.3238
## [13] -2722.2266 -2779.0053 -2825.2022 -2860.4001 -2888.6141 -2897.5927
## [19] -2898.7579 -2897.2001 -2894.0840 -2890.7236 -2884.0979 -2877.1204
## [25] -2869.8948 -2862.6391 -2855.3103 -2847.4223 -2839.3826## [1] 19## (Intercept) Brandasus Brandinfinix Brandoppo BrandRealme
## 11094447.8 -7667292.8 -9972189.6 -8951201.6 -8835170.6
## BrandSamsung Brandvivo Brandxiaomi RAM1 RAM2
## -7169694.4 -8736884.5 -9083600.1 -10723015.7 -7959902.8
## RAM3 RAM4 RAM8 RAM12 RAM16
## -6375916.6 -2763403.3 959749.5 4713574.5 8300264.8
## RAM18 Penyimpanan32 Penyimpanan64 Penyimpanan256 Penyimpanan512
## 12552487.1 3611103.6 -578264.1 495172.0 2434857.9BIC terendah di model 19, Namun sama seperti r-squared dan cp mallows, masih ada variabel dummy yang tidak tertaut ke model
#Kesimpulan
method.sub <- data.frame(kriteria_pemilihan_model = c("R2-adjusted", "Cp", "AIC", "BIC"), model_terpilih = c(24,22, 29, 19))
colnames(method.sub) <- c("Kriteria Pemilihan Model", "Model Terpilih")
method.sub## Kriteria Pemilihan Model Model Terpilih
## 1 R2-adjusted 24
## 2 Cp 22
## 3 AIC 29
## 4 BIC 19Maka dari itu dari Beberapa kriteria pemilihan model, kami memilih model ke 29 karena mengcover seluruh peubah dummy yang ada
#membuat plot
par(mfrow=c(2,2))
#plot Adj R2
plot(reg.summary$adjr2 ,xlab="Model",
ylab="R2-Adjusted",type="l", main="Plot R2-Adjusted")
points(24, reg.summary$adjr2[24], col="red",cex=2,pch=20)
#plot Cp
plot(reg.summary$cp ,xlab="Model",
ylab="Cp",type="l", main="Plot Cp")
points(22, reg.summary$cp[22], col="red",cex=2,pch=20)
#plot AIC
plot(aic, xlab="Model",
ylab="AIC",type="l", main="Plot AIC")
points(29, aic[29], col="red",cex=2,pch=20)
#plot BIC
plot(reg.summary$bic ,xlab="Model",
ylab="BIC",type="l", main="Plot BIC")
points(19, reg.summary$bic[19], col="red",cex=2,pch=20)
Dapat dilihat secara visualisasi bahwa nilai nilai kategori yang kami pakai sesuai dengan plot
Forward
# Model null (hanya intercept)
null_model <- lm(HARGA ~ 1, data = data)
# Model full (semua variabel)
full_model <- lm(HARGA ~ ., data = data)
# Forward selection
forward_model <- step(null_model,
scope = list(lower = null_model, upper = full_model),
direction = "forward")## Start: AIC=95870.92
## HARGA ~ 1
##
## Df Sum of Sq RSS AIC
## + Brand 7 2.2663e+16 4.5677e+16 94627
## + Penyimpanan 6 8.1557e+15 6.0185e+16 95486
## + RAM 9 7.5238e+15 6.0817e+16 95525
## + LOKASI 6 3.3614e+14 6.8004e+16 95868
## <none> 6.8341e+16 95871
## + J.Ulasan 1 1.7083e+12 6.8339e+16 95873
##
## Step: AIC=94627.45
## HARGA ~ Brand
##
## Df Sum of Sq RSS AIC
## + RAM 9 1.8744e+16 2.6933e+16 92997
## + Penyimpanan 6 1.0760e+16 3.4918e+16 93801
## + J.Ulasan 1 1.5941e+14 4.5518e+16 94619
## <none> 4.5677e+16 94627
## + LOKASI 6 1.2717e+14 4.5550e+16 94631
##
## Step: AIC=92996.78
## HARGA ~ Brand + RAM
##
## Df Sum of Sq RSS AIC
## + Penyimpanan 6 1.3425e+15 2.5590e+16 92849
## + J.Ulasan 1 3.7539e+13 2.6895e+16 92994
## <none> 2.6933e+16 92997
## + LOKASI 6 7.6156e+13 2.6857e+16 93000
##
## Step: AIC=92849.2
## HARGA ~ Brand + RAM + Penyimpanan
##
## Df Sum of Sq RSS AIC
## + LOKASI 6 1.1107e+14 2.5479e+16 92848
## + J.Ulasan 1 2.3375e+13 2.5567e+16 92848
## <none> 2.5590e+16 92849
##
## Step: AIC=92847.62
## HARGA ~ Brand + RAM + Penyimpanan + LOKASI
##
## Df Sum of Sq RSS AIC
## <none> 2.5479e+16 92848
## + J.Ulasan 1 1.1932e+13 2.5467e+16 92848##
## Call:
## lm(formula = HARGA ~ Brand + RAM + Penyimpanan + LOKASI, data = data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -12755191 -1607112 -588418 1423541 14641855
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 10981145 164588 66.719 < 2e-16 ***
## Brandasus -7725334 468403 -16.493 < 2e-16 ***
## Brandinfinix -9990208 221399 -45.123 < 2e-16 ***
## Brandoppo -8887418 255200 -34.825 < 2e-16 ***
## BrandRealme -8808185 250139 -35.213 < 2e-16 ***
## BrandSamsung -7177078 173586 -41.346 < 2e-16 ***
## Brandvivo -8714458 197512 -44.121 < 2e-16 ***
## Brandxiaomi -9096452 177503 -51.247 < 2e-16 ***
## RAM1 -11127642 1327145 -8.385 < 2e-16 ***
## RAM2 -8146282 575314 -14.160 < 2e-16 ***
## RAM3 -6411525 361381 -17.742 < 2e-16 ***
## RAM4 -2785169 182153 -15.290 < 2e-16 ***
## RAM8 963907 166128 5.802 7.21e-09 ***
## RAM12 4753831 266839 17.815 < 2e-16 ***
## RAM16 8396291 634543 13.232 < 2e-16 ***
## RAM18 12694810 2099945 6.045 1.67e-09 ***
## RAM24 1562644 1742518 0.897 0.369910
## Penyimpanan8 -3012127 2927824 -1.029 0.303657
## Penyimpanan16 2410139 1154869 2.087 0.036976 *
## Penyimpanan32 3757600 603965 6.222 5.59e-10 ***
## Penyimpanan64 -545256 191560 -2.846 0.004451 **
## Penyimpanan256 517825 137823 3.757 0.000175 ***
## Penyimpanan512 2463878 249320 9.882 < 2e-16 ***
## LOKASIBali 418587 467453 0.895 0.370610
## LOKASIDI Yogyakarta -625319 747780 -0.836 0.403088
## LOKASIJawa Barat 5923 227479 0.026 0.979231
## LOKASIJawa Tengah -158271 477261 -0.332 0.740196
## LOKASIJawa Timur 402264 176339 2.281 0.022604 *
## LOKASILuar Jawa 466372 165410 2.819 0.004841 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 2871000 on 3092 degrees of freedom
## Multiple R-squared: 0.6272, Adjusted R-squared: 0.6238
## F-statistic: 185.8 on 28 and 3092 DF, p-value: < 2.2e-16Dari Model Forward, didapat model dengan 28 peubah , dimana 1 peubah dibuang yaitu peubah jumlah ulasan, Didapat pula nilai r-squared yang cukup tinggi di angka 0.6238
Backward
# Model penuh
full_model <- lm(HARGA ~ ., data = data)
# Backward elimination
backward_model <- step(full_model, direction = "backward")## Start: AIC=92848.16
## HARGA ~ Brand + LOKASI + J.Ulasan + RAM + Penyimpanan
##
## Df Sum of Sq RSS AIC
## - J.Ulasan 1 1.1932e+13 2.5479e+16 92848
## <none> 2.5467e+16 92848
## - LOKASI 6 9.9628e+13 2.5567e+16 92848
## - Penyimpanan 6 1.3624e+15 2.6830e+16 92999
## - RAM 9 9.2964e+15 3.4764e+16 93801
## - Brand 7 3.1323e+16 5.6790e+16 95337
##
## Step: AIC=92847.62
## HARGA ~ Brand + LOKASI + RAM + Penyimpanan
##
## Df Sum of Sq RSS AIC
## <none> 2.5479e+16 92848
## - LOKASI 6 1.1107e+14 2.5590e+16 92849
## - Penyimpanan 6 1.3774e+15 2.6857e+16 93000
## - RAM 9 9.3082e+15 3.4788e+16 93801
## - Brand 7 3.1397e+16 5.6876e+16 95340##
## Call:
## lm(formula = HARGA ~ Brand + LOKASI + RAM + Penyimpanan, data = data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -12755191 -1607112 -588418 1423541 14641855
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 10981145 164588 66.719 < 2e-16 ***
## Brandasus -7725334 468403 -16.493 < 2e-16 ***
## Brandinfinix -9990208 221399 -45.123 < 2e-16 ***
## Brandoppo -8887418 255200 -34.825 < 2e-16 ***
## BrandRealme -8808185 250139 -35.213 < 2e-16 ***
## BrandSamsung -7177078 173586 -41.346 < 2e-16 ***
## Brandvivo -8714458 197512 -44.121 < 2e-16 ***
## Brandxiaomi -9096452 177503 -51.247 < 2e-16 ***
## LOKASIBali 418587 467453 0.895 0.370610
## LOKASIDI Yogyakarta -625319 747780 -0.836 0.403088
## LOKASIJawa Barat 5923 227479 0.026 0.979231
## LOKASIJawa Tengah -158271 477261 -0.332 0.740196
## LOKASIJawa Timur 402264 176339 2.281 0.022604 *
## LOKASILuar Jawa 466372 165410 2.819 0.004841 **
## RAM1 -11127642 1327145 -8.385 < 2e-16 ***
## RAM2 -8146282 575314 -14.160 < 2e-16 ***
## RAM3 -6411525 361381 -17.742 < 2e-16 ***
## RAM4 -2785169 182153 -15.290 < 2e-16 ***
## RAM8 963907 166128 5.802 7.21e-09 ***
## RAM12 4753831 266839 17.815 < 2e-16 ***
## RAM16 8396291 634543 13.232 < 2e-16 ***
## RAM18 12694810 2099945 6.045 1.67e-09 ***
## RAM24 1562644 1742518 0.897 0.369910
## Penyimpanan8 -3012127 2927824 -1.029 0.303657
## Penyimpanan16 2410139 1154869 2.087 0.036976 *
## Penyimpanan32 3757600 603965 6.222 5.59e-10 ***
## Penyimpanan64 -545256 191560 -2.846 0.004451 **
## Penyimpanan256 517825 137823 3.757 0.000175 ***
## Penyimpanan512 2463878 249320 9.882 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 2871000 on 3092 degrees of freedom
## Multiple R-squared: 0.6272, Adjusted R-squared: 0.6238
## F-statistic: 185.8 on 28 and 3092 DF, p-value: < 2.2e-16Dari Model Backward, didapat model dengan 28 peubah , dimana 1 peubah dibuang yaitu peubah jumlah ulasan, Didapat pula nilai r-squared yang cukup tinggi di angka 0.6238
Stepwise
stepwise_model <- step(null_model,
scope = list(lower = null_model, upper = full_model),
direction = "both")## Start: AIC=95870.92
## HARGA ~ 1
##
## Df Sum of Sq RSS AIC
## + Brand 7 2.2663e+16 4.5677e+16 94627
## + Penyimpanan 6 8.1557e+15 6.0185e+16 95486
## + RAM 9 7.5238e+15 6.0817e+16 95525
## + LOKASI 6 3.3614e+14 6.8004e+16 95868
## <none> 6.8341e+16 95871
## + J.Ulasan 1 1.7083e+12 6.8339e+16 95873
##
## Step: AIC=94627.45
## HARGA ~ Brand
##
## Df Sum of Sq RSS AIC
## + RAM 9 1.8744e+16 2.6933e+16 92997
## + Penyimpanan 6 1.0760e+16 3.4918e+16 93801
## + J.Ulasan 1 1.5941e+14 4.5518e+16 94619
## <none> 4.5677e+16 94627
## + LOKASI 6 1.2717e+14 4.5550e+16 94631
## - Brand 7 2.2663e+16 6.8341e+16 95871
##
## Step: AIC=92996.78
## HARGA ~ Brand + RAM
##
## Df Sum of Sq RSS AIC
## + Penyimpanan 6 1.3425e+15 2.5590e+16 92849
## + J.Ulasan 1 3.7539e+13 2.6895e+16 92994
## <none> 2.6933e+16 92997
## + LOKASI 6 7.6156e+13 2.6857e+16 93000
## - RAM 9 1.8744e+16 4.5677e+16 94627
## - Brand 7 3.3884e+16 6.0817e+16 95525
##
## Step: AIC=92849.2
## HARGA ~ Brand + RAM + Penyimpanan
##
## Df Sum of Sq RSS AIC
## + LOKASI 6 1.1107e+14 2.5479e+16 92848
## + J.Ulasan 1 2.3375e+13 2.5567e+16 92848
## <none> 2.5590e+16 92849
## - Penyimpanan 6 1.3425e+15 2.6933e+16 92997
## - RAM 9 9.3273e+15 3.4918e+16 93801
## - Brand 7 3.1773e+16 5.7363e+16 95354
##
## Step: AIC=92847.62
## HARGA ~ Brand + RAM + Penyimpanan + LOKASI
##
## Df Sum of Sq RSS AIC
## <none> 2.5479e+16 92848
## + J.Ulasan 1 1.1932e+13 2.5467e+16 92848
## - LOKASI 6 1.1107e+14 2.5590e+16 92849
## - Penyimpanan 6 1.3774e+15 2.6857e+16 93000
## - RAM 9 9.3082e+15 3.4788e+16 93801
## - Brand 7 3.1397e+16 5.6876e+16 95340##
## Call:
## lm(formula = HARGA ~ Brand + RAM + Penyimpanan + LOKASI, data = data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -12755191 -1607112 -588418 1423541 14641855
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 10981145 164588 66.719 < 2e-16 ***
## Brandasus -7725334 468403 -16.493 < 2e-16 ***
## Brandinfinix -9990208 221399 -45.123 < 2e-16 ***
## Brandoppo -8887418 255200 -34.825 < 2e-16 ***
## BrandRealme -8808185 250139 -35.213 < 2e-16 ***
## BrandSamsung -7177078 173586 -41.346 < 2e-16 ***
## Brandvivo -8714458 197512 -44.121 < 2e-16 ***
## Brandxiaomi -9096452 177503 -51.247 < 2e-16 ***
## RAM1 -11127642 1327145 -8.385 < 2e-16 ***
## RAM2 -8146282 575314 -14.160 < 2e-16 ***
## RAM3 -6411525 361381 -17.742 < 2e-16 ***
## RAM4 -2785169 182153 -15.290 < 2e-16 ***
## RAM8 963907 166128 5.802 7.21e-09 ***
## RAM12 4753831 266839 17.815 < 2e-16 ***
## RAM16 8396291 634543 13.232 < 2e-16 ***
## RAM18 12694810 2099945 6.045 1.67e-09 ***
## RAM24 1562644 1742518 0.897 0.369910
## Penyimpanan8 -3012127 2927824 -1.029 0.303657
## Penyimpanan16 2410139 1154869 2.087 0.036976 *
## Penyimpanan32 3757600 603965 6.222 5.59e-10 ***
## Penyimpanan64 -545256 191560 -2.846 0.004451 **
## Penyimpanan256 517825 137823 3.757 0.000175 ***
## Penyimpanan512 2463878 249320 9.882 < 2e-16 ***
## LOKASIBali 418587 467453 0.895 0.370610
## LOKASIDI Yogyakarta -625319 747780 -0.836 0.403088
## LOKASIJawa Barat 5923 227479 0.026 0.979231
## LOKASIJawa Tengah -158271 477261 -0.332 0.740196
## LOKASIJawa Timur 402264 176339 2.281 0.022604 *
## LOKASILuar Jawa 466372 165410 2.819 0.004841 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 2871000 on 3092 degrees of freedom
## Multiple R-squared: 0.6272, Adjusted R-squared: 0.6238
## F-statistic: 185.8 on 28 and 3092 DF, p-value: < 2.2e-16Dari Model Stepwise, didapat model dengan 28 peubah , dimana 1 peubah dibuang yaitu peubah jumlah ulasan, Didapat pula nilai r-squared yang cukup tinggi di angka 0.6238
library(readxl)
library(tidyverse)
data <- read_excel("C:/Users/Admin/Downloads/PSD Kelompok 3 (3).xlsx")
names(data)[names(data) == "Jumlah Ulasan"] <- "J.Ulasan"
head(data)## # A tibble: 6 × 6
## Brand HARGA LOKASI J.Ulasan RAM Penyimpanan
## <chr> <dbl> <chr> <chr> <chr> <chr>
## 1 iphone 4388000 Jabodetabek 10 4 256
## 2 iphone 9458000 Jabodetabek 3071 4 128
## 3 iphone 9409000 Jabodetabek 2713 4 128
## 4 iphone 5047000 Luar Jawa 87 4 64
## 5 iphone 11287000 Jabodetabek 567 6 128
## 6 iphone 21447000 Jabodetabek 732 8 256data$Brand_iphone <- ifelse(data$Brand == "iphone", 1, 0)
data$Brand_samsung <- ifelse(data$Brand == "samsung", 1, 0)
data$Brand_xiaomi <- ifelse(data$Brand == "xiaomi", 1, 0)
data$J.Ulasan <- as.numeric(data$J.Ulasan)
# Membersihkan data lokasi
data$LOKASI <- gsub("^di ", "", data$LOKASI)
# Membuat dummy DKI Jakarta vs Lainnya
data$dummy <- ifelse(data$LOKASI == "Jabodetabek", 2, 3)
data$dummyram6 <- ifelse(data$RAM == 6, 1, 0)
data$Brand <- relevel(as.factor(data$Brand), ref="iphone")
data$LOKASI <- relevel(as.factor(data$dummy), ref= "2")
data$RAM <- relevel(as.factor(as.numeric(data$RAM)), ref = "1")
data$Penyimpanan <- relevel(as.factor(as.numeric(data$Penyimpanan)), ref = "128")
data$J.Ulasan <- as.numeric(data$J.Ulasan)Pada Tahap ini , kami ingin mencoba re dummy variabel di Lokasi dan RAM dimana kami hanya mempertahankan referencenya saja dan membuat kategori lainnya selain itu dengan dilambangkan dengan angka.
##
## Call:
## lm(formula = HARGA ~ Brand + dummy + dummyram6 + Penyimpanan +
## J.Ulasan, data = data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -11520286 -1620775 -447706 877252 18761573
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 8665481.7 361873.2 23.946 < 2e-16 ***
## Brandasus -2193998.8 459488.7 -4.775 1.88e-06 ***
## Brandinfinix -7863486.7 240782.0 -32.658 < 2e-16 ***
## Brandoppo -6994171.3 287121.2 -24.360 < 2e-16 ***
## BrandRealme -6834263.3 280762.4 -24.342 < 2e-16 ***
## BrandSamsung -4840441.0 178770.5 -27.076 < 2e-16 ***
## Brandvivo -6344122.5 212745.1 -29.820 < 2e-16 ***
## Brandxiaomi -7298562.3 193542.3 -37.710 < 2e-16 ***
## dummy 130104.9 135962.0 0.957 0.33868
## dummyram6 434934.9 172327.7 2.524 0.01166 *
## Penyimpanan8 -8829886.1 3353000.3 -2.633 0.00849 **
## Penyimpanan16 -2875859.9 1273404.2 -2.258 0.02399 *
## Penyimpanan32 -2731171.9 612496.6 -4.459 8.52e-06 ***
## Penyimpanan64 -2791235.0 197462.1 -14.136 < 2e-16 ***
## Penyimpanan256 1690847.0 150751.4 11.216 < 2e-16 ***
## Penyimpanan512 5209691.8 259351.3 20.087 < 2e-16 ***
## J.Ulasan -455.2 326.5 -1.394 0.16340
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 3349000 on 3104 degrees of freedom
## (25 observations deleted due to missingness)
## Multiple R-squared: 0.4907, Adjusted R-squared: 0.4881
## F-statistic: 186.9 on 16 and 3104 DF, p-value: < 2.2e-16Didapat model dengan semua peubah dan Aman, Namun saja r-squarednya sangat kecil dibawah dari metode sebelumnya, sehingga hal ini tidak kami lanjutkan ke pengujian lainnya.