library(readxl)
library(neuralnet)
library(Metrics)
library(readxl)
datainput <- read_excel(path ="neuronetwork.xlsx")
datainput
## # A tibble: 15 x 4
## nilai_rapot nilai_sekolah nilai_ujian_nasional nilai
## <dbl> <dbl> <dbl> <dbl>
## 1 80.6 80.3 41 0.346
## 2 79.3 77.2 41 0.335
## 3 79.3 77.2 41 0.338
## 4 80 77 42 0.342
## 5 79 77 46 0.364
## 6 79 77 40 0.330
## 7 75 76 40 0.318
## 8 80 75 41 0.330
## 9 78 78 39 0.324
## 10 81 79 40 0.337
## 11 80 79 41 0.341
## 12 80 78 41 0.342
## 13 82 80 42 0.353
## 14 83 79 43 0.358
## 15 76 75 39 0.311
jmlBaris <- NROW(datainput)
data_training <- datainput[1:(jmlBaris -2),]
data_training
## # A tibble: 13 x 4
## nilai_rapot nilai_sekolah nilai_ujian_nasional nilai
## <dbl> <dbl> <dbl> <dbl>
## 1 80.6 80.3 41 0.346
## 2 79.3 77.2 41 0.335
## 3 79.3 77.2 41 0.338
## 4 80 77 42 0.342
## 5 79 77 46 0.364
## 6 79 77 40 0.330
## 7 75 76 40 0.318
## 8 80 75 41 0.330
## 9 78 78 39 0.324
## 10 81 79 40 0.337
## 11 80 79 41 0.341
## 12 80 78 41 0.342
## 13 82 80 42 0.353
data_testing <- datainput[(jmlBaris-2) :jmlBaris,]
data_testing
## # A tibble: 3 x 4
## nilai_rapot nilai_sekolah nilai_ujian_nasional nilai
## <dbl> <dbl> <dbl> <dbl>
## 1 82 80 42 0.353
## 2 83 79 43 0.358
## 3 76 75 39 0.311
nnmodel <- neuralnet(nilai ~ nilai_rapot + nilai_sekolah, data = data_training, hidden = c(3), linear.output = T)
plot(nnmodel)
result_nn <- compute(nnmodel, data_testing[,2:3])
result_nn[["net.result"]]
## [,1]
## [1,] 0.3387687
## [2,] 0.3387687
## [3,] 0.3387687
errornn = rmse(data_testing$nilai, result_nn$net.result)
errornn
## [1] 0.02126721