r - Error in plot.window(...) : need finite 'xlim' values

Question

What should i do for this error? My code is :

library(e1071)
library(hydroGOF)
donnees <- read.csv("F:/new work with shahab/Code-SVR/SVR/MainData.csv")
summary(donnees)

#partitioning into training and testing set
donnees.train <- donnees[donnees$subset=="train",2:ncol(donnees)]
donnees.test <- donnees[donnees$subset=="test",2:ncol(donnees)]

#use the mean of the dependent variable as a predictor
def.pred <- mean(donnees.train$y)

#error sum of squares of the default model on the test set
def.rss <- sum((donnees.test$y-def.pred)^2)
print(def.rss)
plot(donnees.train)
#*****************
#linear regression
#*****************
#Linear Models
reg <- lm(y ~., data = donnees.train)
print(summary(reg))
#error sum of squares of the model on the test set
reg.pred <- predict(reg,newdata = donnees.test)
reg.rss <- sum((donnees.test$y-reg.pred)^2)
print(reg.rss)

#pseudo-r-squared
print(1.0-reg.rss/def.rss)


#**********************************
#rbf epsilon-svr with cost = 1.0
#**********************************
epsilon.svr <- svm(y ~.,data = donnees.train, scale = T, type = "eps-regression",
                   kernel = "radial", cost = 1.0, epsilon=0.1,tolerance=0.001, shrinking=T,
                   fitted=T)
print(epsilon.svr)
#prédiction
esvr.pred <- predict(epsilon.svr,newdata = donnees.test)
esvr.rss <- sum((donnees.test$y-esvr.pred)^2)
#pseudo-R2
print(1.0-esvr.rss/def.rss)
esvr.rmse=rmse(donnees.test$y,esvr.pred)
print(esvr.rmse)

#****************************************************
#detect the "best" cost parameter for rbf epsilon-svr
#****************************************************
costs <- seq(from=0.05,to=3.0,by=0.005)
pseudor2 <- double(length(costs))
for (c in 1:length(costs)){
  epsilon.svr <- svm(y ~.,data = donnees.train, scale = T, type = "eps-regression",
                     kernel = "radial", cost = costs[c], epsilon=0.1,tolerance=0.001, shrinking=T,
                     fitted=T)
  #prédiction
  esvr.pred <- predict(epsilon.svr,newdata = donnees.test)
  esvr.rss <- sum((donnees.test$y-esvr.pred)^2)
  pseudor2[c] <- 1.0-esvr.rss/def.rss
}

#graphical representation
plot(costs,pseudor2,type="l")
#show the max. of pseudo-r2 and the corresponding cost parameter
print(max(pseudor2))
k <- which.max(pseudor2)
print(costs[k])

And my maindata in excel worksheet is :

    subset  x1  x2  y       
train   18  1088    9.77        
train   0   831 5.96        
train   0   785 5.36        
train   0   762 5.08        
train   0   749 4.92        
train   0.5 731 4.69        
train   0   727 4.64        
train   2   743 4.84        
train   5   818 5.83        
train   12  942 7.49        
train   13  973 7.98        
train   89.5    1292    12.94       
train   46.5    1086    9.61        
train   5.5 877 6.59        
train   1   826 5.89        
train   0.5 780 5.3     
train   3.5 756 5       
train   4   764 5.1     
train   28.5    851 6.26        
train   10  866 6.45        
train   20.5    839 6.09        
train   7   759 5.03        
train   0.5 722 4.57        
train   0   708 4.4     
train   0   694 4.22        
train   0   689 4.16        
train   0   679 4.03        
train   11  769 5.2     
train   0.5 697 4.26        
train   10.5    702 4.33        
train   1.5 692 4.2     
train   3   743 4.86        
train   16  958 7.98        
train   14  835 6.05        
train   0   713 4.46        
train   0.5 671 3.94        
train   0   659 3.79        
train   0   646 3.63        
train   0.5 636 3.52        
train   0   627 3.43        
train   0   629 3.44        
train   1   682 4.1     
train   8.5 735 4.81        
train   1   729 4.67        
train   0   649 3.66        
train   56  774 5.29        
train   1.5 663 3.84        
train   5.5 787 5.49        
train   50  839 6.14        
train   6.5 699 4.29        
train   1.5 756 5.03        
train   11.5    669 3.91        
train   5   684 4.1     
train   0   653 3.71        
train   0.5 669 3.94        
train   0   638 3.53        
train   0.5 647 3.65        
train   12.5    715 4.56        
train   7.5 921 7.37        
train   50  1149    10.95       
train   10.5    772 5.21        
train   23.5    1205    11.93       
train   23.5    1171    11.01       
train   8.5 927 7.26        
train   0.5 1009    8.45        
train   4   1019    8.62        
train   0   968 7.88        
train   2   862 6.38        
train   22  1349    14.15       
train   16.5    1029    8.74        
train   8.5 846 6.15        
train   0.5 853 6.26        
train   9.5 819 5.81        
train   19.5    775 5.24        
train   23  746 4.88        
train   46.5    723 4.58        
train   1   733 4.72        
train   26.5    731 4.69        
train   34.5    814 5.81        
train   2   743 4.84        
train   0   715 4.49        
train   4   680 4.05        
train   8   816 5.85        
train   20  823 5.91        
train   0.5 824 5.93        
train   2.5 746 4.88        
train   0   817 5.87        
train   0   732 4.7     
train   6   682 4.07        
train   0   685 4.12        
train   1   719 4.56        
train   10.5    701 4.31        
train   23.5    1002    8.74        
train   23.5    947 7.71        
train   8.5 808 5.66        
train   0.5 835 6.06        
train   4   811 5.71        
train   0   709 4.42        
train   2   696 4.25        
train   22  913 7.21        
train   16.5    860 6.42        
train   8.5 902 7.15        
train   0.5 781 5.32        
train   9.5 862 6.45        
train   19.5    833 6.02        
train   23  803 5.63        
train   46.5    903 7.06        
train   1   822 5.86        
train   26.5    1040    9.19        
train   34.5    939 7.55        
train   2   793 5.48        
train   0   730 4.68        
train   4   719 4.53        
train   8   706 4.38        
train   20  829 5.99        
train   0.5 724 4.6     
train   2.5 697 4.26        
train   0   669 3.91        
train   0   657 3.76        
train   6   724 4.66        
train   0   657 3.76        
train   1   676 4.02        
train   23.5    968 8.24        
train   0   696 4.25        
train   12  727 4.73        
train   0.5 651 3.69        
train   3.5 685 4.12        
train   0.5 668 3.9     
train   0   626 3.4     
train   0   619 3.32        
train   1   697 4.34        
train   0.5 624 3.37        
train   13.5    683 4.14        
train   0   651 3.68        
train   0   621 3.33        
train   0   612 3.24        
train   3   668 3.91        
train   0   626 3.39        
train   0.5 614 3.27        
train   0   614 3.26        
train   2.5 630 3.45        
train   0.5 617 3.3     
train   0   616 3.3     
train   8   684 4.14        
train   0.5 612 3.24        
train   0   598 3.09        
train   0   588 2.99        
train   0   590 3       
train   6   648 3.71        
train   0   598 3.1     
train   2   614 3.29        
train   33  804 5.9     
train   0   619 3.32        
train   0   588 2.98        
train   0   577 2.87        
train   0   571 2.81        
train   0.5 572 2.82        
train   4.5 607 3.2     
train   0   579 2.89        
train   0   562 2.72        
train   0   565 2.74        
train   0   554 2.63        
train   0   543 2.51        
train   0   536 2.44        
train   0   531 2.39        
train   0   532 2.4     
train   0.5 529 2.36        
train   0   527 2.35        
train   0   528 2.36        
train   0   523 2.31        
train   0   521 2.29        
train   0   523 2.31        
train   0.5 541 2.49        
train   0   522 2.3     
train   0.5 533 2.42        
train   2   529 2.37        
train   10  638 3.65        
train   0.5 544 2.52        
train   5   627 3.52        
train   0   535 2.43        
train   0   516 2.24        
train   0   520 2.27        
train   32  841 6.55        
train   11.5    838 6.29        
train   0   595 3.06        
train   0.5 592 3.03        
train   0   558 2.67        
train   0   540 2.48        
train   0   534 2.42        
train   2   539 2.46        
train   13  623 3.42        
train   0   553 2.62        
train   0   561 2.71        
train   0   546 2.55        
train   0   512 2.2     
train   2   518 2.26        
train   32  702 4.46        
train   27  731 4.76        
train   1   604 3.15        
train   0   584 2.94        
train   0   548 2.57        
train   0   519 2.26        
train   29.5    735 4.91        
train   0   564 2.74        
train   12  606 3.23        
train   0   542 2.51        
train   0   516 2.24        
train   0   508 2.15        
train   0   500 2.07        
train   0   495 2.03        
train   0   496 2.04        
train   0   492 1.99        
train   0   496 2.04        
train   0   490 1.98        
train   0   494 2.02        
train   0   490 1.99        
train   3   548 2.62        
train   17  546 2.61        
train   9.5 737 4.95        
train   1.5 584 2.96        
train   0   521 2.27        
train   0.5 526 2.34        
train   0   539 2.48        
train   24.5    699 4.45        
train   41  740 4.97        
train   3   569 2.8     
train   1   525 2.32        
train   0   511 2.18        
train   0   498 2.05        
train   2   597 3.22        
train   0.5 520 2.27        
train   66  909 7.77        
train   23  716 4.54        
train   0.5 564 2.74        
train   4.5 582 2.94        
train   0   577 2.88        
train   0   527 2.34        
train   0   512 2.19        
train   0   503 2.09        
train   8.5 561 2.73        
train   0   533 2.4     
train   24.5    640 3.77        
train   0   515 2.21        
train   0   496 2.03        
train   0   485 1.93        
train   0   480 1.88        
train   0   476 1.85        
train   0   480 1.88        
train   24  689 4.34        
train   0   568 2.79        
train   0   506 2.12        
train   8.5 680 4.19        
train   12  657 3.87        
train   5.5 635 3.61        
train   19.5    761 5.18        
train   1.5 567 2.77        
train   3.5 678 4.1     
train   4   574 2.84        
train   7   628 3.5     
train   6   656 3.77        
train   0   551 2.6     
train   0.5 526 2.33        
train   0.5 555 2.64        
train   8.5 666 4.01        
train   1   564 2.74        
train   0   534 2.41        
train   0   521 2.27        
train   7.5 599 3.15        
train   4.5 585 2.96        
train   3   647 3.65        
train   0   547 2.56        
train   0   531 2.38        
train   0   508 2.15        
train   0   500 2.08        
train   0   503 2.09        
train   0   492 1.99        
train   0.5 492 1.99        
train   5   647 3.92        
train   0   513 2.19        
train   6.5 523 2.3     
train   2   527 2.35        
train   2   522 2.3     
train   22.5    817 6.14        
train   18.5    808 5.86        
train   8.5 775 5.37        
train   4.5 705 4.37        
train   58  891 6.96        
train   7   642 3.58        
train   7   614 3.29        
train   10.5    772 5.29        
train   7.5 714 4.54        
train   3.5 613 3.25        
train   6   575 2.85        
train   24.5    680 4.19        
train   18.5    801 5.64        
train   0   640 3.55        
train   6.5 610 3.23        
train   0.5 592 3.03        
train   36.5    835 6.2     
test    0   673 3.97    2.97    2.49
test    0.5 571 2.81    3.74    2.3
test    0   553 2.62    3.56    3.1
test    6   597 3.17    3.52    3.46
test    7   584 2.97    3.75    3.6
test    4.5 649 3.74    3.76    3.5
test    9.5 636 3.56    5.27    5.4
test    14.5    629 3.52    3.69    3.65
test    6.5 648 3.75    

Answer 1

The problem is that you're (probably) trying to plot a vector that consists exclusively of missing (NA) values. Here's an example:

> x=rep(NA,100)
> y=rnorm(100)
> plot(x,y)
Error in plot.window(...) : need finite 'xlim' values
In addition: Warning messages:
1: In min(x) : no non-missing arguments to min; returning Inf
2: In max(x) : no non-missing arguments to max; returning -Inf

In your example this means that in your line plot(costs,pseudor2,type="l"), costs is completely NA. You have to figure out why this is, but that's the explanation of your error.

---

From comments:

Scott C Wilson: Another possible cause of this message (not in this case, but in others) is attempting to use character values as X or Y data. You can use the class function to check your x and Y values to be sure if you think this might be your issue.

stevec: Here is a quick and easy solution to that problem (basically wrap x in as.factor(x))