Differences

This shows you the differences between two versions of the page.

--- r:neural_network [2016/12/14 08:22] – [E.G. 2] hkimscil
+++ r:neural_network [2016/12/14 09:02] (current) – [E.G 5] hkimscil
@@ Line 502: / Line 502: @@
 > paste("오차율 =", round(error.overall, digit = 2), "%")
 [1] "오차율 = 1.85 %"</code>
+====== E.G. 3 ======
+{{:r:creditset.csv}}
+<WRAP info 70%>The dataset contains information on different clients who received a loan at least 10 years ago. The variables income (yearly), age, loan (size in euros) and LTI (the loan to yearly income ratio) are available. Our goal is to devise a model which predicts, based on the input variables LTI and age, whether or not a default will occur within 10 years.
+</WRAP>
+<code>set.seed(1234567890)
+library("neuralnet")
+dataset <- read.csv("creditset.csv")
+head(dataset)
+# extract a set to train the NN
+trainset <- dataset[1:800, ]
+# select the test set
+testset <- dataset[801:2000, ]
+## build the neural network (NN)
+creditnet <- neuralnet(default10yr ~ LTI + age, trainset, hidden = 4, lifesign = "minimal",
+    linear.output = FALSE, threshold = 0.1)
+## plot the NN
+plot(creditnet, rep = "best")
+</code>
+__test the resulting output__
+<code>temp_test <- subset(testset, select = c("LTI", "age"))
+creditnet.results <- compute(creditnet, temp_test)
+head(temp_test)
+results <- data.frame(actual = testset$default10yr, prediction = creditnet.results$net.result)
+results[100:115, ]
+results$prediction <- round(results$prediction)
+results[100:115, ]
+pred.table <- table(testset$default10yr, results$prediction)
+pred.table
+library(plyr)
+count(testset, 'default10yr')
+</code>
+with outputs
+<code>> set.seed(1234567890)
+> library("neuralnet")
+>
+> dataset <- read.csv("creditset.csv")
+> head(dataset)
+  clientid      income         age          loan             LTI
+        1 66155.92510 59.01701507 8106.53213129 0.1225367511623
+        2 34415.15397 48.11715310 6564.74501768 0.1907515806612
+        3 57317.17006 63.10804949 8020.95329639 0.1399397996720
+        4 42709.53420 45.75197235 6103.64226014 0.1429105321411
+        5 66952.68885 18.58433593 8770.09923520 0.1309894999955
+        6 24904.06414 57.47160710   15.49859844 0.0006223320961
+  default10yr
+           0
+           0
+           0
+           0
+           1
+           0
+>
+> # extract a set to train the NN
+> trainset <- dataset[1:800, ]
+>
+> # select the test set
+> testset <- dataset[801:2000, ]
+>
+> ## build the neural network (NN)
+> creditnet <- neuralnet(default10yr ~ LTI + age, trainset, hidden = 4, lifesign = "minimal",
++                        linear.output = FALSE, threshold = 0.1)
+hidden: 4    thresh: 0.1    rep: 1/1    steps:    7266	error: 0.79202	time: 2.94 secs
+>
+> ## plot the NN
+> plot(creditnet, rep = "best")
+>
+</code>
+<code>> temp_test <- subset(testset, select = c("LTI", "age"))
+>
+> creditnet.results <- compute(creditnet, temp_test)
+> head(temp_test)
+              LTI         age
+0.02306808811 25.90644520
+0.13729704954 40.77430558
+0.10456984914 32.47350580
+0.15985046411 53.22813215
+0.11161429579 46.47915325
+0.11489364221 47.12736998
+>
+> results <- data.frame(actual = testset$default10yr, prediction = creditnet.results$net.result)
+> results[100:115, ]
+    actual                                 prediction
+      0 0.0000000000000000000000000015964854322398
+      0 0.0000000000000000000000000065162871249459
+      0 0.0000000000164043993271687692878796349660
+      1 0.9999999999219191249011373656685464084148
+      0 0.0000000000000000013810778585990655628959
+      0 0.0000000000000000539636283549268839978413
+      0 0.0000000000000000000234592312583964807452
+      1 0.9581419934268182725389806364546529948711
+      0 0.2499229633059938393557786184828728437424
+      0 0.0000000000000007044361454974903653282470
+      0 0.0006082559674722681341413332845036165963
+      1 0.9999999878713862200285689141310285776854
+      0 0.0000000000000000000000000015562211243506
+      1 0.9999999993455563895849991240538656711578
+      0 0.0000000000000000000000000000003082538282
+      0 0.0000000019359618836434052080615331181690
+>
+> results$prediction <- round(results$prediction)
+> results[100:115, ]
+    actual prediction
+      0          0
+      0          0
+      0          0
+      1          1
+      0          0
+      0          0
+      0          0
+      1          1
+      0          0
+      0          0
+      0          0
+      1          1
+      0          0
+      1          1
+      0          0
+      0          0
+</code>
+<code>> pred.table <- table(testset$default10yr, results$prediction)
+> pred.table
+    1
+1035    1
+    3  161
+>
+> library(plyr)
+> count(testset, 'default10yr')
+  default10yr freq
+           0 1036
+           1  164
+</code>
+====== E.G 5 ======