白戸則道®#'@ rm(list = ls(all = TRUE))
suppressPackageStartupMessages(library('BBmisc'))
pkgs <- c('knitr', 'kableExtra', 'tint', 'devtools', 'lubridate', 'data.table', 'quantmod', 'shiny', 'R.utils', 'memoise', 'magrittr', 'plyr', 'dplyr', 'purrr', 'broom', 'sparklyr', 'rsparkling', 'h2o', 'stringr', 'tidyverse', 'htmltools', 'highcharter', 'googleVis', 'formattable', 'DT', 'fst', 'googleVis')
suppressAll(lib(pkgs))
## Set option to below if you want to plot an independent webpage with graph
#'@ op <- options(gvis.plot.tag=NULL)
op <- options(gvis.plot.tag = 'chart')
options(warn = -1)
#'@ options(rsparkling.sparklingwater.version = '2.2.0')
#'@ spark_install(version = '2.2.0')
#'@ sc <- spark_connect(master = 'local', version = '2.2.0')
rm(pkgs)1. Introduction
Initially, I would like to write a shiny app1 for the dictionary and user data search engine, but due to I saw the file size of the dataset will cause a heavily loading, then I forget about shiny app and started write this html web page.
Basically, this paper analyse the bank risk and loan quality. This is an assessment for joinning a company named Money Lion as a data scientist.
In paper Credit Scoring in R2, the author compare logistic regression, random forest and random forest-logistic regression and finally concludes that later model is best fitted.
2. Data
Due to the high volume data, here I use data.table::fread to speed up the reading time.
## Read `loan.csv`
loan <- fread('./data/loan.csv') %>% tbl_df %>% mutate(applicationDate = str_replace_all(applicationDate, 'T', ' ') %>% ymd_hms, originatedDate = str_replace_all(originatedDate, 'T', ' ') %>% ymd_hms) %>% mutate_if(is.character, funs(factor))##
Read 31.2% of 577682 rows
Read 51.9% of 577682 rows
Read 60.6% of 577682 rows
Read 577682 rows and 19 (of 19) columns from 0.102 GB file in 00:00:06## Read `payment.csv`
payment <- fread('./data/payment.csv') %>% tbl_df %>% mutate(paymentDate = str_replace_all(paymentDate, 'T', ' ') %>% ymd_hms) %>% mutate_if(is.character, funs(factor))##
Read 95.7% of 689364 rows
Read 689364 rows and 9 (of 9) columns from 0.050 GB file in 00:00:03## Read `clarity_underwriting_variables.csv`
cuv <- fread('./data/clarity_underwriting_variables.csv') %>% tbl_df %>% mutate_if(is.character, funs(factor))Here I try to merge both datasets loan and payment and arrange the data.
loan_payment <- plyr::join(payment, loan, by = 'loanId') %>% tbl_df %>% mutate(paymentDuration = difftime(paymentDate, lag(paymentDate), units = 'days'))Sparklyr
Besides, I try to using sparklyr3 for high volume data modelling due to it only spend few seconds time to model and predict moreover compare among models.
## copy data to spark
#'@ loan_tbl <- copy_to(sc, loan, 'loan')
#'@ payment_tbl <- copy_to(sc, payment, 'payment')
#'@ cuv_tbl <- copy_to(sc, cuv, 'cuv')
#'@ loan_payment_tbl <- copy_to(sc, loan_payment, 'loan_payment')Data Visualization
The pie chart below shows the portion loan status of customers.
hchart(as.character(loan$loanStatus), type = 'pie')graph 2.1 : loan status
Well, below chart shows the occupation of payment status onto the approved loan.
hchart(as.character(loan_payment$paymentStatus), type = 'pie')graph 2.2 : payment status
hchart(loan$loanAmount)graph 2.3 : loan amount
Below histogram shows the installment index, which is the term of installment.
hchart(payment$installmentIndex)graph 2.4 : installment term
3. Modelling
> ##sample data for credit score modelling.
> head(loan_data)
loan_status loan_amnt int_rate grade emp_length home_ownership annual_inc age
1 0 5000 10.65 B 10 RENT 24000 33
2 0 2400 NA C 25 RENT 12252 31
3 0 10000 13.49 C 13 RENT 49200 24
4 0 5000 NA A 3 RENT 36000 39
5 0 3000 NA E 9 RENT 48000 24
6 0 12000 12.69 B 11 OWN 75000 28source : introduction-to-credit-risk-modeling-in-r
I don’t pretend to know the correct model, here I try to build some logistic regression to get the best fit.
## Due to errors in MCMCpack::MCMClogit() and biglm::bigglm(), here I am using glm().
#summary(lm1 <- lm(Fertility ~ ., data = swiss))
#slm1 <- step(lm1)
#summary(slm1)
#slm1$anova
## loan
loan1 <- loan %>% dplyr::select(-c(loanId, anon_ssn, applicationDate, originatedDate, clarityFraudId))
fit.loan <- glm(formula = loanStatus ~ ., family = binomial(link = logit), data = loan1)
#'@ glm(formula = loanStatus ~ payFrequency + apr + originated + nPaidOff + loanAmount + originallyScheduledPaymentAmount + state + leadType + leadCost + hasCF, family = binomial(link = logit), data = loan1)
sfit.loan <- step(fit.loan)## Start: AIC=1985.6
## loanStatus ~ payFrequency + apr + originated + nPaidOff + approved +
## isFunded + loanAmount + originallyScheduledPaymentAmount +
## state + leadType + leadCost + fpStatus + hasCF
##
## Df Deviance AIC
## - fpStatus 7 1840.1 1976.1
## - isFunded 1 1835.6 1983.6
## - approved 1 1835.6 1983.6
## - originallyScheduledPaymentAmount 1 1837.2 1985.2
## <none> 1835.6 1985.6
## - loanAmount 1 1837.6 1985.6
## - originated 1 1844.3 1992.3
## - payFrequency 5 1855.0 1995.0
## - apr 1 1848.1 1996.1
## - nPaidOff 1 1848.7 1996.7
## - leadCost 1 1860.0 2008.0
## - hasCF 1 1882.1 2030.1
## - state 44 2056.5 2118.5
## - leadType 9 2080.7 2212.7
##
## Step: AIC=1976.1
## loanStatus ~ payFrequency + apr + originated + nPaidOff + approved +
## isFunded + loanAmount + originallyScheduledPaymentAmount +
## state + leadType + leadCost + hasCF
##
## Df Deviance AIC
## - isFunded 1 1840.1 1974.1
## - approved 1 1840.1 1974.1
## <none> 1840.1 1976.1
## - originallyScheduledPaymentAmount 1 1847.9 1981.9
## - loanAmount 1 1848.0 1982.0
## - originated 1 1848.9 1982.9
## - payFrequency 5 1859.1 1985.1
## - apr 1 1854.2 1988.2
## - nPaidOff 1 1854.6 1988.6
## - leadCost 1 1864.8 1998.8
## - hasCF 1 1886.4 2020.4
## - state 44 2060.1 2108.1
## - leadType 9 2085.6 2203.6
##
## Step: AIC=1974.1
## loanStatus ~ payFrequency + apr + originated + nPaidOff + approved +
## loanAmount + originallyScheduledPaymentAmount + state + leadType +
## leadCost + hasCF
##
## Df Deviance AIC
## - approved 1 1840.1 1972.1
## <none> 1840.1 1974.1
## - originallyScheduledPaymentAmount 1 1847.9 1979.9
## - loanAmount 1 1848.0 1980.0
## - originated 1 1848.9 1980.9
## - payFrequency 5 1859.1 1983.1
## - apr 1 1854.2 1986.2
## - nPaidOff 1 1854.6 1986.6
## - leadCost 1 1864.8 1996.8
## - hasCF 1 1886.4 2018.4
## - state 44 2060.1 2106.1
## - leadType 9 2085.6 2201.6
##
## Step: AIC=1972.1
## loanStatus ~ payFrequency + apr + originated + nPaidOff + loanAmount +
## originallyScheduledPaymentAmount + state + leadType + leadCost +
## hasCF
##
## Df Deviance AIC
## <none> 1840.1 1972.1
## - originallyScheduledPaymentAmount 1 1847.9 1977.9
## - loanAmount 1 1848.0 1978.0
## - payFrequency 5 1859.1 1981.1
## - apr 1 1854.2 1984.2
## - nPaidOff 1 1854.6 1984.6
## - leadCost 1 1864.8 1994.8
## - hasCF 1 1886.4 2016.4
## - originated 1 1891.8 2021.8
## - state 44 2060.1 2104.1
## - leadType 9 2085.6 2199.6summary(sfit.loan)##
## Call:
## glm(formula = loanStatus ~ payFrequency + apr + originated +
## nPaidOff + loanAmount + originallyScheduledPaymentAmount +
## state + leadType + leadCost + hasCF, family = binomial(link = logit),
## data = loan1)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -4.6215 0.0000 0.0043 0.0113 1.0958
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 3.747e+00 1.717e+04 0.000 0.99983
## payFrequencyB 8.781e-01 5.068e+03 0.000 0.99986
## payFrequencyI 1.477e+01 5.166e+03 0.003 0.99772
## payFrequencyM 6.675e-01 5.068e+03 0.000 0.99989
## payFrequencyS 1.889e-01 5.068e+03 0.000 0.99997
## payFrequencyW 7.662e-02 5.068e+03 0.000 0.99999
## apr 8.092e-03 1.730e-03 4.677 2.91e-06
## originatedTrue 1.777e+01 7.755e+02 0.023 0.98172
## nPaidOff 3.154e+01 8.470e+02 0.037 0.97030
## loanAmount 3.000e-03 1.091e-03 2.750 0.00595
## originallyScheduledPaymentAmount -8.660e-04 3.146e-04 -2.753 0.00591
## stateAK -3.494e+00 1.651e+04 0.000 0.99983
## stateAL -2.069e+00 1.651e+04 0.000 0.99990
## stateAZ -3.460e+00 1.651e+04 0.000 0.99983
## stateCA 1.752e+01 1.652e+04 0.001 0.99915
## stateCO 1.902e+01 1.684e+04 0.001 0.99910
## stateCT -3.191e+00 1.651e+04 0.000 0.99985
## stateDE -2.691e+00 1.651e+04 0.000 0.99987
## stateFL -2.000e+00 1.651e+04 0.000 0.99990
## stateGA -1.158e+00 1.651e+04 0.000 0.99994
## stateHI -2.941e+00 1.651e+04 0.000 0.99986
## stateIA -2.845e+00 1.651e+04 0.000 0.99986
## stateID 1.596e+01 1.802e+04 0.001 0.99929
## stateIL 1.315e+00 1.651e+04 0.000 0.99994
## stateIN -1.841e+00 1.651e+04 0.000 0.99991
## stateKS -3.623e+00 1.651e+04 0.000 0.99982
## stateKY -3.875e+00 1.651e+04 0.000 0.99981
## stateLA -1.258e+00 1.651e+04 0.000 0.99994
## stateMD -1.912e+01 2.210e+05 0.000 0.99993
## stateMI -9.931e-01 1.651e+04 0.000 0.99995
## stateMN -2.750e+00 1.651e+04 0.000 0.99987
## stateMO 1.529e+00 1.651e+04 0.000 0.99993
## stateMS -3.426e+00 1.651e+04 0.000 0.99983
## stateNC 1.504e+01 1.655e+04 0.001 0.99927
## stateND -4.276e+00 1.651e+04 0.000 0.99979
## stateNE 1.616e+01 1.819e+04 0.001 0.99929
## stateNJ -2.027e+00 1.651e+04 0.000 0.99990
## stateNM -1.963e+00 1.651e+04 0.000 0.99991
## stateNV -4.565e-01 1.651e+04 0.000 0.99998
## stateNY -3.755e+00 2.210e+05 0.000 0.99999
## stateOH 4.199e-01 1.651e+04 0.000 0.99998
## stateOH-TEST 1.474e+01 2.166e+05 0.000 0.99995
## stateOK -3.347e+00 1.651e+04 0.000 0.99984
## statePA -2.532e+00 1.651e+04 0.000 0.99988
## stateRI 1.580e+01 1.811e+04 0.001 0.99930
## stateSC -6.257e-02 1.651e+04 0.000 1.00000
## stateSD -2.124e+00 1.651e+04 0.000 0.99990
## stateTN 1.538e+01 1.655e+04 0.001 0.99926
## stateTX 1.520e-02 1.651e+04 0.000 1.00000
## stateTX-TEST -3.055e+01 2.166e+05 0.000 0.99989
## stateUT -5.610e-01 1.651e+04 0.000 0.99997
## stateVA 1.770e+01 1.664e+04 0.001 0.99915
## stateWA -2.480e+00 1.651e+04 0.000 0.99988
## stateWI 1.363e+00 1.651e+04 0.000 0.99993
## stateWY 1.550e+01 1.713e+04 0.001 0.99928
## leadTypecalifornia -2.803e+00 6.162e+03 0.000 0.99964
## leadTypeexpress -8.086e+01 2.512e+04 -0.003 0.99743
## leadTypeinstant-offer -2.700e+01 5.527e+02 -0.049 0.96104
## leadTypelead -5.022e+00 4.312e-01 -11.647 < 2e-16
## leadTypelionpay 2.176e+01 4.391e+04 0.000 0.99960
## leadTypeorganic 1.389e+01 1.107e+03 0.013 0.98999
## leadTypeprescreen 1.308e+01 2.380e+03 0.005 0.99561
## leadTyperc_returning -1.719e+01 2.944e+03 -0.006 0.99534
## leadTyperepeat -1.719e+01 3.419e+04 -0.001 0.99960
## leadCost 4.688e-02 1.026e-02 4.571 4.86e-06
## hasCF 2.497e+00 4.295e-01 5.815 6.07e-09
##
## (Intercept)
## payFrequencyB
## payFrequencyI
## payFrequencyM
## payFrequencyS
## payFrequencyW
## apr ***
## originatedTrue
## nPaidOff
## loanAmount **
## originallyScheduledPaymentAmount **
## stateAK
## stateAL
## stateAZ
## stateCA
## stateCO
## stateCT
## stateDE
## stateFL
## stateGA
## stateHI
## stateIA
## stateID
## stateIL
## stateIN
## stateKS
## stateKY
## stateLA
## stateMD
## stateMI
## stateMN
## stateMO
## stateMS
## stateNC
## stateND
## stateNE
## stateNJ
## stateNM
## stateNV
## stateNY
## stateOH
## stateOH-TEST
## stateOK
## statePA
## stateRI
## stateSC
## stateSD
## stateTN
## stateTX
## stateTX-TEST
## stateUT
## stateVA
## stateWA
## stateWI
## stateWY
## leadTypecalifornia
## leadTypeexpress
## leadTypeinstant-offer
## leadTypelead ***
## leadTypelionpay
## leadTypeorganic
## leadTypeprescreen
## leadTyperc_returning
## leadTyperepeat
## leadCost ***
## hasCF ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 2525.7 on 573706 degrees of freedom
## Residual deviance: 1840.1 on 573641 degrees of freedom
## (3975 observations deleted due to missingness)
## AIC: 1972.1
##
## Number of Fisher Scoring iterations: 24sfit.loan$anovaBasically, for loan quality modelling will use customers’ detail like employment status, income, commiment etc., however I used available dataset loan + payment for the modelling here.
## loan payment
loan_payment1 <- loan_payment %>% dplyr::select(-c(loanId, anon_ssn, applicationDate, originatedDate, paymentDate, clarityFraudId, paymentReturnCode))
fit.loanpayment <- glm(formula = paymentStatus ~ ., family = binomial(link = logit), data = loan_payment1)
sfit.loanpayment <- step(fit.loanpayment)## Start: AIC=570069.1
## paymentStatus ~ installmentIndex + isCollection + principal +
## fees + paymentAmount + payFrequency + apr + originated +
## nPaidOff + approved + isFunded + loanStatus + loanAmount +
## originallyScheduledPaymentAmount + state + leadType + leadCost +
## fpStatus + hasCF + paymentDuration
##
##
## Step: AIC=573532.2
## paymentStatus ~ installmentIndex + isCollection + principal +
## fees + paymentAmount + payFrequency + apr + originated +
## nPaidOff + approved + isFunded + loanStatus + loanAmount +
## originallyScheduledPaymentAmount + state + leadType + leadCost +
## fpStatus + paymentDurationsummary(sfit.loanpayment)##
## Call:
## glm(formula = paymentStatus ~ installmentIndex + isCollection +
## principal + fees + paymentAmount + payFrequency + apr + originated +
## nPaidOff + approved + isFunded + loanStatus + loanAmount +
## originallyScheduledPaymentAmount + state + leadType + leadCost +
## fpStatus + paymentDuration, family = binomial(link = logit),
## data = loan_payment1)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -5.3394 -0.7586 0.0339 0.7706 4.5557
##
## Coefficients:
## Estimate Std. Error z value
## (Intercept) 8.367e+07 4.249e+10 0.002
## installmentIndex -1.004e-01 6.306e-04 -159.183
## isCollectionTRUE 3.135e-01 2.223e-02 14.106
## principal -3.052e-03 5.994e-04 -5.092
## fees -1.611e-03 6.057e-04 -2.660
## paymentAmount 5.312e-03 5.940e-04 8.942
## payFrequencyI 6.108e-02 3.008e-02 2.031
## payFrequencyM -1.622e-01 2.095e-02 -7.744
## payFrequencyS -2.944e-01 1.592e-02 -18.495
## payFrequencyW 5.131e-01 8.635e-03 59.420
## apr -5.494e-03 6.429e-05 -85.452
## originatedTrue 1.216e+01 6.444e+01 0.189
## nPaidOff -2.980e-01 5.573e-03 -53.485
## approvedTrue -1.658e+01 6.444e+01 -0.257
## isFunded -8.367e+07 4.249e+10 -0.002
## loanStatusCharged Off Paid Off 3.711e-01 9.784e-01 0.379
## loanStatusCredit Return Void -8.367e+07 4.249e+10 -0.002
## loanStatusCSR Voided New Loan -8.367e+07 4.249e+10 -0.002
## loanStatusCustomer Voided New Loan -8.367e+07 4.249e+10 -0.002
## loanStatusCustomver Voided New Loan -8.367e+07 4.249e+10 -0.002
## loanStatusExternal Collection 1.174e+00 9.771e-01 1.201
## loanStatusInternal Collection 2.521e+00 9.771e-01 2.580
## loanStatusNew Loan 9.338e+00 9.804e-01 9.524
## loanStatusPaid Off Loan 2.350e+00 9.771e-01 2.405
## loanStatusPending Paid Off 4.852e+00 9.802e-01 4.950
## loanStatusPending Rescind -8.367e+07 4.249e+10 -0.002
## loanStatusReturned Item 9.721e+00 1.001e+00 9.713
## loanStatusSettled Bankruptcy 1.152e+00 9.776e-01 1.179
## loanStatusSettlement Paid Off 1.872e+00 9.773e-01 1.916
## loanStatusSettlement Pending Paid Off 1.463e+01 1.394e+02 0.105
## loanStatusVoided New Loan -8.367e+07 4.249e+10 -0.002
## loanStatusWithdrawn Application -8.367e+07 4.249e+10 -0.002
## loanAmount -1.026e-03 3.306e-05 -31.031
## originallyScheduledPaymentAmount 6.921e-05 1.132e-05 6.115
## stateAL 2.254e-02 1.857e-01 0.121
## stateAZ 5.886e-02 1.822e-01 0.323
## stateCA 5.463e-01 1.809e-01 3.020
## stateCO -7.962e-01 1.896e-01 -4.198
## stateCT -6.496e-02 1.841e-01 -0.353
## stateDE -1.028e-01 1.994e-01 -0.516
## stateFL 2.452e-01 1.804e-01 1.359
## stateGA 1.588e+00 1.898e-01 8.363
## stateHI 2.180e-01 2.194e-01 0.993
## stateIA 3.454e-01 1.907e-01 1.811
## stateID -1.649e-01 2.179e-01 -0.757
## stateIL 1.393e-01 1.805e-01 0.772
## stateIN 9.441e-01 1.803e-01 5.236
## stateKS 1.643e-02 1.892e-01 0.087
## stateKY 1.587e-01 1.843e-01 0.861
## stateLA 1.469e-01 1.861e-01 0.789
## stateMD 1.143e+01 1.034e+02 0.111
## stateMI 8.601e-01 1.803e-01 4.770
## stateMN -1.138e-01 1.852e-01 -0.615
## stateMO 7.193e-01 1.803e-01 3.989
## stateMS -1.032e-01 1.875e-01 -0.551
## stateNC 1.134e+00 1.804e-01 6.289
## stateND 1.439e-01 2.241e-01 0.642
## stateNE -4.955e-02 1.994e-01 -0.249
## stateNJ 4.055e-01 1.809e-01 2.242
## stateNM 9.199e-01 1.876e-01 4.902
## stateNV 2.954e-01 1.818e-01 1.625
## stateOH 1.177e+00 1.799e-01 6.544
## stateOK 5.315e-03 1.912e-01 0.028
## statePA 2.748e-01 1.809e-01 1.519
## stateRI 1.755e-01 2.029e-01 0.865
## stateSC 1.620e+00 1.808e-01 8.962
## stateSD 3.993e-01 1.932e-01 2.067
## stateTN 6.777e-01 1.805e-01 3.755
## stateTX 1.428e+00 1.802e-01 7.927
## stateUT 4.825e-01 1.860e-01 2.594
## stateVA -1.074e-01 1.821e-01 -0.590
## stateWA 1.071e-01 1.862e-01 0.575
## stateWI 4.769e-01 1.803e-01 2.645
## stateWY 7.957e-01 1.935e-01 4.113
## leadTypecalifornia -9.916e-01 7.895e-02 -12.559
## leadTypeexpress 3.253e+00 2.353e-01 13.822
## leadTypeinstant-offer 1.217e+01 2.483e+01 0.490
## leadTypelead -4.470e-02 1.162e-02 -3.848
## leadTypelionpay -4.139e-01 3.825e-01 -1.082
## leadTypeorganic 1.513e-02 1.070e-02 1.413
## leadTypeprescreen -3.724e-01 2.231e-02 -16.692
## leadTyperc_returning 1.525e-01 2.598e-02 5.870
## leadTyperepeat 8.757e-01 1.982e-01 4.418
## leadCost 3.892e-03 2.410e-04 16.146
## fpStatusCancelled -1.570e+01 3.902e+01 -0.402
## fpStatusChecked -1.289e+01 3.902e+01 -0.330
## fpStatusPending -6.909e+00 9.409e+01 -0.073
## fpStatusRejected -1.387e+01 3.902e+01 -0.355
## fpStatusReturned -1.527e+01 3.903e+01 -0.391
## fpStatusSkipped -1.323e+01 3.902e+01 -0.339
## paymentDuration -2.230e-02 2.431e-04 -91.710
## Pr(>|z|)
## (Intercept) 0.998429
## installmentIndex < 2e-16 ***
## isCollectionTRUE < 2e-16 ***
## principal 3.55e-07 ***
## fees 0.007804 **
## paymentAmount < 2e-16 ***
## payFrequencyI 0.042280 *
## payFrequencyM 9.61e-15 ***
## payFrequencyS < 2e-16 ***
## payFrequencyW < 2e-16 ***
## apr < 2e-16 ***
## originatedTrue 0.850331
## nPaidOff < 2e-16 ***
## approvedTrue 0.796932
## isFunded 0.998429
## loanStatusCharged Off Paid Off 0.704487
## loanStatusCredit Return Void 0.998429
## loanStatusCSR Voided New Loan 0.998429
## loanStatusCustomer Voided New Loan 0.998429
## loanStatusCustomver Voided New Loan 0.998429
## loanStatusExternal Collection 0.229639
## loanStatusInternal Collection 0.009874 **
## loanStatusNew Loan < 2e-16 ***
## loanStatusPaid Off Loan 0.016154 *
## loanStatusPending Paid Off 7.41e-07 ***
## loanStatusPending Rescind 0.998429
## loanStatusReturned Item < 2e-16 ***
## loanStatusSettled Bankruptcy 0.238457
## loanStatusSettlement Paid Off 0.055373 .
## loanStatusSettlement Pending Paid Off 0.916393
## loanStatusVoided New Loan 0.998429
## loanStatusWithdrawn Application 0.998429
## loanAmount < 2e-16 ***
## originallyScheduledPaymentAmount 9.68e-10 ***
## stateAL 0.903393
## stateAZ 0.746618
## stateCA 0.002525 **
## stateCO 2.69e-05 ***
## stateCT 0.724199
## stateDE 0.605939
## stateFL 0.174110
## stateGA < 2e-16 ***
## stateHI 0.320471
## stateIA 0.070132 .
## stateID 0.449163
## stateIL 0.440361
## stateIN 1.64e-07 ***
## stateKS 0.930811
## stateKY 0.389179
## stateLA 0.429911
## stateMD 0.911993
## stateMI 1.84e-06 ***
## stateMN 0.538759
## stateMO 6.65e-05 ***
## stateMS 0.581902
## stateNC 3.20e-10 ***
## stateND 0.520782
## stateNE 0.803746
## stateNJ 0.024969 *
## stateNM 9.46e-07 ***
## stateNV 0.104221
## stateOH 5.99e-11 ***
## stateOK 0.977823
## statePA 0.128749
## stateRI 0.387203
## stateSC < 2e-16 ***
## stateSD 0.038712 *
## stateTN 0.000173 ***
## stateTX 2.24e-15 ***
## stateUT 0.009479 **
## stateVA 0.555457
## stateWA 0.565097
## stateWI 0.008177 **
## stateWY 3.91e-05 ***
## leadTypecalifornia < 2e-16 ***
## leadTypeexpress < 2e-16 ***
## leadTypeinstant-offer 0.623955
## leadTypelead 0.000119 ***
## leadTypelionpay 0.279238
## leadTypeorganic 0.157593
## leadTypeprescreen < 2e-16 ***
## leadTyperc_returning 4.36e-09 ***
## leadTyperepeat 9.96e-06 ***
## leadCost < 2e-16 ***
## fpStatusCancelled 0.687391
## fpStatusChecked 0.741130
## fpStatusPending 0.941465
## fpStatusRejected 0.722308
## fpStatusReturned 0.695545
## fpStatusSkipped 0.734579
## paymentDuration < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 922836 on 688980 degrees of freedom
## Residual deviance: 573350 on 688890 degrees of freedom
## (383 observations deleted due to missingness)
## AIC: 573532
##
## Number of Fisher Scoring iterations: 12sfit.loanpayment$anovaFrom above models, we take the lowest AIC value as best fit model for this study.
4. Conclusion
I’ll need to learn and realise the business of credit risk in order to .
5. Appendix
5.1 Documenting File Creation
It’s useful to record some information about how your file was created.
- File creation date: 2017-10-25
- File latest updated date: 2017-10-28
- R version 3.4.2 (2017-09-28)
- R version (short form): 3.4.2
- rmarkdown package version: 1.6.0.9004
- tufte package version: 0.2
- File version: 1.0.1
- Author Profile: ®γσ, Eng Lian Hu
- GitHub: Source Code
- Additional session information
[1] “2017-10-28 20:39:53 JST”
5.2 Reference
- A simple amortization function
- Level payment loan functions in R
- Mortgage Calculator (and Amortization Charts) with R
- r package :
tvm - r package :
optiRum - Guide to Credit Scoring in R
- The
lifecontingenciesPackage - Performing Financial and Actuarial Mathematics Calculations in R
Powered by - Copyright® Intellectual Property Rights of 
Scibrokes®個人の経営企業
similar with my previous project in Coursera Data Science Capstone which search from dictionary and display the predicted risk.↩
Kindly refer to paper 6th in reference section.↩
You can refer to Sparklyr: Using Spark with R Markdown for more information.↩