Commit ef7ee0b6 authored by Marius Puke's avatar Marius Puke
Browse files

add naive example

parent 7daf25c1

.gitignore

+2 −1
Original line number Diff line number Diff line 
01_data_raw/



amex_outcomes.rds

amex_outcomes.csv

.Rproj.user

amex_submission_naive_forecasters.csv

0 → 100644
+10001 −0

File added.

Preview size limit exceeded, changes collapsed.

data_prep.R

+5 −0
Original line number Diff line number Diff line
@@ -50,3 +50,8 @@ submission <- validation %>% select(ID) %>% mutate(PD = NA) 
saveRDS(submission,"amex_submission.rds")

write.csv(submission, "amex_submission.csv")



# correct classes

outcomes <- datraw_small[20001:30000,] %>% select(ID, target)

outcomes

saveRDS(outcomes,"amex_outcomes.rds")

write.csv(outcomes, "amex_outcomes.csv")

naive_example.R

+46 −19
Original line number Diff line number Diff line 
rm(list=ls())

setwd(dirname(rstudioapi::getActiveDocumentContext()$path))

library(tidyverse)

# Local path to save the file

url_train <- "https://aidaho-edu.uni-hohenheim.de/gitlab/mariuspuke/sl2024_classificationcompetition/-/raw/main/amex_train.csv?ref_type=heads"

train <- read.csv(url_validation)





library(httr)

response <- GET(url_submission)

response$url





readRDS(rawConnection(content(response, "raw", encoding = "ISO-8859-1")))



readRDS(rawConnection(content(response, "raw")))



readRDS(rawConnection(content(submission, "raw")))



# Local path to save the file

destination <- "amex_submission_dl.rds"

# train 

url_train <- "https://aidaho-edu.uni-hohenheim.de/gitlab/mariuspuke/sl2024_classificationcompetition/-/raw/main/amex_train.csv?ref_type=heads"

train <- read.csv(url_train)%>% tibble() %>% select(-X)

train



url_validation <- "https://aidaho-edu.uni-hohenheim.de/gitlab/mariuspuke/sl2024_classificationcompetition/-/raw/main/amex_validation.csv?ref_type=heads"

validation <- read.csv(url_validation) %>% tibble() %>% select(-X)

validation



url_submission <- "https://aidaho-edu.uni-hohenheim.de/gitlab/mariuspuke/sl2024_classificationcompetition/-/raw/main/amex_submission.csv?ref_type=heads"

submission <- read.csv(url_submission) %>% tibble() %>% select(-X)

submission





# fit a naive lm 

catvars <- c('B_30', 'B_38', 'D_114', 'D_116', 'D_117', 'D_120', 'D_126', 'D_63', 'D_64',  'D_68')

lmod <- lm(target ~ ., data = train %>% select(target, catvars))

lmod



# probability of default in validation set

pd <- 

validation %>% 

  mutate(

    #probability of default

    pd = predict(lmod, newdata = validation)

  ) %>% 

  select(ID,pd)



final <- 

left_join(

  submission, pd

  ) %>% 

  mutate(PD=pd) %>% 

  select(ID,PD)

write.csv(final, "amex_submission_naive_forecasters.csv")





outcomes <- readRDS("amex_outcomes.rds") %>% 

  left_join(

  .,final

)



library(pROC)

roc_curve <- roc(as.numeric(outcomes$target), outcomes$PD)

roc_curve

plot(roc_curve, main="ROC Curve", col="#1c61b6")

# Download the file

download.file(url, destination, mode = "wb")
 
 No newline at end of file