Klasifikasi Data Pengguna Kartu Kredit dengan Metode Suport Vector Machine
Hellan Bagus Sedyansyah
28 Februari 2020
Support Vector Machine (SVM) diperkenalkan oleh Vapnik pada tahun 1992 sebagai suatu teknik klasifikasi yang efisien untuk masalah nonlinier. SVM berbeda dengan teknik klasifikasi di era 1980-an, seperti decission tree dan ANN, yang secara konsep kurang begitu jelas dan seringkali terjebak pada optimum lokal. SVM memiliki konsep yang jauh lebih matang, lebih jelas secara matematis, dibanding teknik-teknik klasifikasi sebelum era 1990-an. (Suyanto,2018)
Membaca data
bank <- read.csv("UniversalBank.csv")Struktur Bank
str(bank)## 'data.frame': 5000 obs. of 14 variables:
## $ ID : int 1 2 3 4 5 6 7 8 9 10 ...
## $ Age : int 25 45 39 35 35 37 53 50 35 34 ...
## $ Experience : int 1 19 15 9 8 13 27 24 10 9 ...
## $ Income : int 49 34 11 100 45 29 72 22 81 180 ...
## $ ZIP.Code : int 91107 90089 94720 94112 91330 92121 91711 93943 90089 93023 ...
## $ Family : int 4 3 1 1 4 4 2 1 3 1 ...
## $ CCAvg : num 1.6 1.5 1 2.7 1 0.4 1.5 0.3 0.6 8.9 ...
## $ Education : int 1 1 1 2 2 2 2 3 2 3 ...
## $ Mortgage : int 0 0 0 0 0 155 0 0 104 0 ...
## $ Personal.Loan : int 0 0 0 0 0 0 0 0 0 1 ...
## $ Securities.Account: int 1 1 0 0 0 0 0 0 0 0 ...
## $ CD.Account : int 0 0 0 0 0 0 0 0 0 0 ...
## $ Online : int 0 0 0 0 0 1 1 0 1 0 ...
## $ CreditCard : int 0 0 0 0 1 0 0 1 0 0 ...Deskripsi statistik data
library(psych)
describe(bank)## vars n mean sd median trimmed mad
## ID 1 5000 2500.50 1443.52 2500.5 2500.50 1853.25
## Age 2 5000 45.34 11.46 45.0 45.38 14.83
## Experience 3 5000 20.10 11.47 20.0 20.13 14.83
## Income 4 5000 73.77 46.03 64.0 68.83 43.00
## ZIP.Code 5 5000 93152.50 2121.85 93437.0 93236.13 1968.89
## Family 6 5000 2.40 1.15 2.0 2.37 1.48
## CCAvg 7 5000 1.94 1.75 1.5 1.65 1.33
## Education 8 5000 1.88 0.84 2.0 1.85 1.48
## Mortgage 9 5000 56.50 101.71 0.0 32.98 0.00
## Personal.Loan 10 5000 0.10 0.29 0.0 0.00 0.00
## Securities.Account 11 5000 0.10 0.31 0.0 0.01 0.00
## CD.Account 12 5000 0.06 0.24 0.0 0.00 0.00
## Online 13 5000 0.60 0.49 1.0 0.62 0.00
## CreditCard 14 5000 0.29 0.46 0.0 0.24 0.00
## min max range skew kurtosis se
## ID 1 5000 4999 0.00 -1.20 20.41
## Age 23 67 44 -0.03 -1.15 0.16
## Experience -3 43 46 -0.03 -1.12 0.16
## Income 8 224 216 0.84 -0.05 0.65
## ZIP.Code 9307 96651 87344 -12.49 485.52 30.01
## Family 1 4 3 0.16 -1.40 0.02
## CCAvg 0 10 10 1.60 2.64 0.02
## Education 1 3 2 0.23 -1.55 0.01
## Mortgage 0 635 635 2.10 4.75 1.44
## Personal.Loan 0 1 1 2.74 5.52 0.00
## Securities.Account 0 1 1 2.59 4.69 0.00
## CD.Account 0 1 1 3.69 11.61 0.00
## Online 0 1 1 -0.39 -1.84 0.01
## CreditCard 0 1 1 0.90 -1.18 0.01Analisis data tak bernilai
library(Amelia)## Loading required package: Rcpp## Warning: package 'Rcpp' was built under R version 3.5.3## ##
## ## Amelia II: Multiple Imputation
## ## (Version 1.7.5, built: 2018-05-07)
## ## Copyright (C) 2005-2020 James Honaker, Gary King and Matthew Blackwell
## ## Refer to http://gking.harvard.edu/amelia/ for more information
## ##missmap(bank, main="Missings Map", col=c("yellow", "black"), legend=FALSE)
#Data tidak mengandung NAMerapikan data untuk persiapan pembuatan data training dan testing
bank=bank[,-c(1,5)]
bank$CreditCard <- as.factor(bank$CreditCard)
Under=ifelse(bank$Education==1,1,0)
Grad=ifelse(bank$Education== 2,1,0)
bank=cbind(bank,Under,Grad)
bank_2=bank[,-6]
bank_2[-11] <- scale(bank_2[-11])Membuat data traing dan testing
library(caTools)
set.seed(1111)
spl_1 = sample.split(bank_2$CreditCard, SplitRatio=0.7)
Training = subset(bank_2, spl_1==TRUE)
Testing = subset(bank_2, spl_1==T)Membuat model SVM
library(e1071)
models <- svm(CreditCard~., data=Training, scale=T, kernel="linear", type="C-classification" )
summary(models)##
## Call:
## svm(formula = CreditCard ~ ., data = Training, kernel = "linear",
## type = "C-classification", scale = T)
##
##
## Parameters:
## SVM-Type: C-classification
## SVM-Kernel: linear
## cost: 1
## gamma: 0.08333333
##
## Number of Support Vectors: 2016
##
## ( 1067 949 )
##
##
## Number of Classes: 2
##
## Levels:
## 0 1Pengujian data testing
pred = predict(models, Testing[-11])
tabel <- table(as.numeric(pred),as.numeric(Testing$CreditCard))
library(caret)## Warning: package 'caret' was built under R version 3.5.3## Loading required package: lattice## Loading required package: ggplot2## Warning: package 'ggplot2' was built under R version 3.5.3##
## Attaching package: 'ggplot2'## The following objects are masked from 'package:psych':
##
## %+%, alphaconfusionMatrix(tabel)## Confusion Matrix and Statistics
##
##
## 1 2
## 1 2425 863
## 2 46 166
##
## Accuracy : 0.7403
## 95% CI : (0.7254, 0.7547)
## No Information Rate : 0.706
## P-Value [Acc > NIR] : 3.651e-06
##
## Kappa : 0.1857
##
## Mcnemar's Test P-Value : < 2.2e-16
##
## Sensitivity : 0.9814
## Specificity : 0.1613
## Pos Pred Value : 0.7375
## Neg Pred Value : 0.7830
## Prevalence : 0.7060
## Detection Rate : 0.6929
## Detection Prevalence : 0.9394
## Balanced Accuracy : 0.5714
##
## 'Positive' Class : 1
##