Scribd Logo
Upload

Welcome to Scribd!

  • 23 views

    No3 Uas

    Uploaded by

    Putri Ramatillah Ramadhana
    1. The document performs linear discriminant analysis (LDA) on data to classify observations into two groups (Merk 1 and Merk 2) based on three independent variables. 2. The results show that the prior probabilities of the two groups are equal, and the group means are different for all three variables. 3. A confusion matrix is generated to evaluate the accuracy of the LDA classification, showing that 61.1% of observations from group 1 and 75% from group 2 are correctly classified.

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd

    No3 Uas

    Uploaded by

    Putri Ramatillah Ramadhana
    23 views7 pages
    1. The document performs linear discriminant analysis (LDA) on data to classify observations into two groups (Merk 1 and Merk 2) based on three independent variables. 2. The results show that the prior probabilities of the two groups are equal, and the group means are different for all three variables. 3. A confusion matrix is generated to evaluate the accuracy of the LDA classification, showing that 61.1% of observations from group 1 and 75% from group 2 are correctly classified.

    Original Title

    no3_uas

    Copyright

    © © All Rights Reserved

    Available Formats

    DOCX, PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    1. The document performs linear discriminant analysis (LDA) on data to classify observations into two groups (Merk 1 and Merk 2) based on three independent variables. 2. The results show that the prior probabilities of the two groups are equal, and the group means are different for all three variables. 3. A confusion matrix is generated to evaluate the accuracy of the LDA classification, showing that 61.1% of observations from group 1 and 75% from group 2 are correctly classified.

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd
    Download as docx, pdf, or txt
    23 views7 pages

    No3 Uas

    Uploaded by

    Putri Ramatillah Ramadhana
    1. The document performs linear discriminant analysis (LDA) on data to classify observations into two groups (Merk 1 and Merk 2) based on three independent variables. 2. The results show that the prior probabilities of the two groups are equal, and the group means are different for all three variables. 3. A confusion matrix is generated to evaluate the accuracy of the LDA classification, showing that 61.1% of observations from group 1 and 75% from group 2 are correctly classified.

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd
    Download as docx, pdf, or txt
    You are on page 1of 7

    no3_uas

    Rianti Isnania

    2022-12-15
    library(readxl)
    data = read_excel("D:/KULIAH/SEMESTER 5/APG/UAS/A_R STUDIO
    BISMILLAH/data_diskriminan.xlsx")
    View(data)
    library(MASS)
    fit = lda(Merk~efisiensi+detik+kilo,data=data)
    fit

    ## Call:
    ## lda(Merk ~ efisiensi + detik + kilo, data = data)
    ##
    ## Prior probabilities of groups:
    ## 1 2
    ## 0.5294118 0.4705882
    ##
    ## Group means:
    ## efisiensi detik kilo
    ## 1 15.11111 9.388889 850.3333
    ## 2 14.62500 10.875000 817.4375
    ##
    ## Coefficients of linear discriminants:
    ## LD1
    ## efisiensi -0.568427144
    ## detik 0.719818449
    ## kilo 0.008766566

    Prior probabilities of groups: 1 2 0.5 0.5 prior prob adalah pel 1 obs untuk masuk dalam
    suatu kelompok sblm data terkumpul
    Coefficients of linear discriminants: LD1 y1 -0.13987912 y2 -0.42340040 y3 0.04247445
    y4 -0.02878734 y5 -0.39508971 y6 -0.16819799
    fungsi diskriminan z = -0.14 y1 - 0.423 y2 + 0.04 y3 - 0.03 y4 - 0.395 y5 - 0.168 y6
    #UJI ASUMSI MULTIVARIATE NORMAL #uji multivariate normal
    library(MVN)
    result <- mvn(data =data[2:4],mvnTest = "mardia")
    result

    ## $multivariateNormality
    ## Test Statistic p value Result
    ## 1 Mardia Skewness 14.0411066594509 0.171124850424771 YES
    ## 2 Mardia Kurtosis -0.726011433817873 0.467831761224173 YES
    ## 3 MVN <NA> <NA> YES
    ##
    ## $univariateNormality
    ## Test Variable Statistic p value Normality
    ## 1 Anderson-Darling efisiensi 0.5963 0.1124 YES
    ## 2 Anderson-Darling detik 1.0119 0.0100 NO
    ## 3 Anderson-Darling kilo 1.1592 0.0043 NO
    ##
    ## $Descriptives
    ## n Mean Std.Dev Median Min Max 25th 75th
    Skew
    ## efisiensi 34 14.88235 1.647122 15.0 12.0 19.0 14.0 16.00
    0.3399752
    ## detik 34 10.08824 2.108678 11.0 5.0 13.0 9.0 11.75 -
    0.7147725
    ## kilo 34 834.85294 165.203119 787.3 639.6 1190.8 701.1 965.15
    0.6832758
    ## Kurtosis
    ## efisiensi -0.3672187
    ## detik -0.3869701
    ## kilo -0.8406343

    library(MVN) # Uji multivariate normal jika skala datanya numerik


    mvn(data = data[,c(2:4)], multivariatePlot = "qq") #6 KRN VARIABEL ADA 6
    ## $multivariateNormality
    ## Test HZ p value MVN
    ## 1 Henze-Zirkler 1.046217 0.009117251 NO
    ##
    ## $univariateNormality
    ## Test Variable Statistic p value Normality
    ## 1 Anderson-Darling efisiensi 0.5963 0.1124 YES
    ## 2 Anderson-Darling detik 1.0119 0.0100 NO
    ## 3 Anderson-Darling kilo 1.1592 0.0043 NO
    ##
    ## $Descriptives
    ## n Mean Std.Dev Median Min Max 25th 75th
    Skew
    ## efisiensi 34 14.88235 1.647122 15.0 12.0 19.0 14.0 16.00
    0.3399752
    ## detik 34 10.08824 2.108678 11.0 5.0 13.0 9.0 11.75 -
    0.7147725
    ## kilo 34 834.85294 165.203119 787.3 639.6 1190.8 701.1 965.15
    0.6832758
    ## Kurtosis
    ## efisiensi -0.3672187
    ## detik -0.3869701
    ## kilo -0.8406343

    Dari grafik Chi-Square QQ Plot diatas, dapat dilihat bahwasanya secara umum terbentuk
    garis linear X = Y, maka dapat dikatakan bahwa data berdistribusi multivariate normal.
    #matriks ragam peragam antar kategori sama
    library(biotools)

    ## ---
    ## biotools version 4.2

    #melakukan uji statistik Box’s M di R dapat menggunakan fungsi boxM.


    boxM(data=data[,c(2:4)],grouping=data$Merk)

    ##
    ## Box's M-test for Homogeneity of Covariance Matrices
    ##
    ## data: data[, c(2:4)]
    ## Chi-Sq (approx.) = 8.2156, df = 6, p-value = 0.2227

    Output diatas menunjukkan bahwa dengan tingkat signifikansi 5%, didapat keputusan
    untuk gagal tolak hipotesis nol sehingga tidak terdapat perbedaan matriks ragam
    peragamnya dan dapat digunakan model diskriminan linear
    Terdapat perbedaan rata-rata antar kategori H0 : Tidak terdapat perbedaan rata-rata skor
    diskriminan antar kelompok secara multivariat H1 : terdapat perbedaan rata-rata skor
    diskriminan antar kelompok secara multivariat
    (wilk <- manova(formula = cbind(data$efisiensi, data$detik,
    data$kilo)~data$Merk))

    ## Call:
    ## manova(formula = cbind(data$efisiensi, data$detik, data$kilo) ~
    ## data$Merk)
    ##
    ## Terms:
    ## data$Merk Residuals
    ## resp 1 2.0 87.5
    ## resp 2 18.7 128.0
    ## resp 3 9166.3 891472.0
    ## Deg. of Freedom 1 32
    ##
    ## Residual standard errors: 1.653857 2.000217 166.9087
    ## Estimated effects may be unbalanced

    summary(object = wilk, test = "Wilks")

    ## Df Wilks approx F num Df den Df Pr(>F)


    ## data$Merk 1 0.77154 2.9611 3 30 0.04799 *
    ## Residuals 32
    ## ---
    ## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

    otput diatas menunjukkan bahwa dengan tingkat signifikansi 5%, didapat keputusan untuk
    menolak hipotesis nol atau dengan kata lain terdapat perbedaan rata-rata (nilai variabel
    prediktor) antar kategori
    #Correlation dan multikolinearity check
    kor<-cor(data[,2:4])
    kor

    ## efisiensi detik kilo


    ## efisiensi 1.0000000 -0.3022847 0.7211341
    ## detik -0.3022847 1.0000000 -0.6652169
    ## kilo 0.7211341 -0.6652169 1.0000000

    canonical discriminant analysis (analisis diskriminan kanonik) untuk melihat hubungan


    antara variabel diskriminan dengan variabel independen secara multivariat
    library(candisc)

    ## Loading required package: car

    ## Loading required package: carData

    ## Loading required package: heplots

    ## Loading required package: broom


    ##
    ## Attaching package: 'heplots'

    ## The following object is masked from 'package:biotools':


    ##
    ## boxM

    ##
    ## Attaching package: 'candisc'

    ## The following object is masked from 'package:stats':


    ##
    ## cancor

    #Canonical Correlation (CC)


    cc<-candisc(wilk)
    cc

    ##
    ## Canonical Discriminant Analysis for data$Merk:
    ##
    ## CanRsq Eigenvalue Difference Percent Cumulative
    ## 1 0.22846 0.29611 100 100
    ##
    ## Test of H0: The canonical correlations in the
    ## current row and all that follow are zero
    ##
    ## LR test stat approx F numDF denDF Pr(> F)
    ## 1 0.77154 2.9611 3 30 0.04799 *
    ## ---
    ## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

    #Eigenvalue menunjukan ada tidaknya multikolinearitas antar variabel. Didapatkan nilai


    Eigenvalue = 0.14993. Dikarenakan nilai eigenvalue mendekati 0, maka terjadi
    multikolinearitas antar variabel independen.
    #Canonical Corellation Canonical Correlation (CC)=0.13038, artinya bahwa hubungan
    antara variabel diskriminan dengan variabel independen secara multivariat sebesar
    0.13038 atau besarnya (CC)²= (0.13038)²
    #Jadi dapat disimpulkan bahwa (0.13038)² * 100% variasi antara merk m1 dan m2 dapat
    dijelaskan oleh variabel X1, x2,x3.
    library(dplyr)

    ##
    ## Attaching package: 'dplyr'

    ## The following object is masked from 'package:car':


    ##
    ## recode
    ## The following object is masked from 'package:MASS':
    ##
    ## select

    ## The following objects are masked from 'package:stats':


    ##
    ## filter, lag

    ## The following objects are masked from 'package:base':


    ##
    ## intersect, setdiff, setequal, union

    m1 = data %>%filter(Merk==1) %>% select(efisiensi,detik,kilo)

    m2 = data %>%filter(Merk==2) %>% select(efisiensi,detik,kilo)

    s1=cov(m1)
    s2= cov(m2)
    Spooled<-function(s1,s2,n1,n2){
    W1<-(n1-1)*s1
    W2<-(n2-1)*s2
    Sp<-(1/((n1-1)+(n2-1)))*(W1+W2)
    return(Sp)
    }
    Spl<-Spooled(s1,s2,18,16)
    Spl

    ## efisiensi detik kilo


    ## efisiensi 2.7352431 -0.8914931 198.1268
    ## detik -0.8914931 4.0008681 -226.0362
    ## kilo 198.1268229 -226.0361979 27858.4999

    xbar1 = matrix(c(15.118,9.647,842.188))
    xbar2 = matrix(c(14.647,10.529,827.518))

    a=t(xbar1-xbar2)%*%solve(Spl)

    xbar.1 = a%*%xbar1
    xbar.2 = a%*%xbar2

    #cutting score
    m.kep = 0.5*(xbar.1+xbar.2)
    m.kep

    ## [,1]
    ## [1,] -3.581862

    #untuk melihat akurasi pred klasifikasi dengan confussion matrix

    fit.val = predict(fit,data[,-1]) #keluarkan kolom groupnya


    ct = table(data$Merk, fit.val$class)
    ct

    ##
    ## 1 2
    ## 1 11 7
    ## 2 4 12

    1 2

    150205
    5 observasi dengan kelompok asal 1 diklasifikasikan benar ke kelompok 1
    5 observasi dengan kelompok asal 1 diklasifikasikan benar ke kelompok 1
    0 observasi dengan kelompok asal 1 salah diklasifikasikan ke kelompok 2
    0 observasi dengan kelompok asal 2 salah diklasifikasikan ke kelompok 1
    #membuat confussion matrix untuk melihat akurasi prediksi klasifikasi
    diag(prop.table(ct,1))

    ## 1 2
    ## 0.6111111 0.7500000

    Interpretasi : 1 2 1 1
    100 persen kelompok 1 diklasifikasikan dengan benar. 100 persen kel 2 diklasifikasikan
    dengan benar.
    #total persen correct
    sum(diag(prop.table(ct)))

    ## [1] 0.6764706

    1 interpretasi : 100 persen data diklasifikasikan dengan benar.

    You might also like

    pFad - Phonifier reborn

    Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

    Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


    Alternative Proxies:

    Alternative Proxy

    pFad Proxy

    pFad v3 Proxy

    pFad v4 Proxy