CODEWITHSUNDEEP
r
houda benhar
houda benhar

Reputation: 101

Error in UseMethod("SK") : no applicable method for 'SK' applied to an object of class "character"

I'm using the ScottKnott package to perform some analysis on some data using the following code:

 Data <- read.csv('hf dataset.csv',header=TRUE)
 sk1 <- with(Data,
            SK(x=Data$classifier,
               y=Data$kappa,
               model='y ~ x',
               which='x'))

The code was working just fine. This is an example of the output I used to have using the code above: enter image description here

BUT now I'm getting this

Error in UseMethod("SK") : no applicable method for 'SK' applied to an object of class "character"

using the SAME code and SAME data. This is a snippet of Data:

str(Data)
'data.frame':   1200 obs. of  4 variables:
 $ classifier: chr  "U-KNNCFS" "U-KNNCFS" "U-KNNCFS" "U-KNNCFS" ...
 $ kappa     : num  0.524 0.831 0.561 0.4 0.615 ...
 $ acc       : num  80 93.3 83.3 73.3 83.3 ...
 $ auc       : num  0.762 0.889 0.754 0.7 0.8 ...

I converted the Data$classifier column from "chr" to "Factor" but I still had the same error with "Factor" instead of "character". Can someone help? I've been trying to figure out the problem for 3 days but I'm stuck here.

Here is a small sample of Data:

structure(list(classifier = c("U-KNNCFS", "U-KNNCFS", "U-KNNCFS", 
"U-KNNCFS", "U-KNNCFS", "U-KNNCFS", "U-KNNCFS", "U-KNNCFS", "U-KNNCFS", 
"U-KNNCFS", "U-KNNCON", "U-KNNCON", "U-KNNCON", "U-KNNCON", "U-KNNCON", 
"U-KNNCON", "U-KNNCON", "U-KNNCON", "U-KNNCON", "U-KNNCON", "U-KNNIG4", 
"U-KNNIG4", "U-KNNIG4", "U-KNNIG4", "U-KNNIG4", "U-KNNIG4", "U-KNNIG4", 
"U-KNNIG4", "U-KNNIG4", "U-KNNIG4"), kappa = c(0.523809524, 0.830508475, 
0.561403509, 0.4, 0.615384615, -0.216216216, 0.7, 0.368421053, 
0.7, 0.449367089, 0.307692308, 0.150943396, 0.206349206, 0.052631579, 
0.210526316, -0.054054054, 0.368421053, 0.108108108, 0.307692308, 
-0.080229226, 0.523809524, 0.830508475, 0.561403509, 0.4, 0.615384615, 
-0.216216216, 0.7, 0.368421053, 0.7, 0.449367089), acc= c(80, 
93.33333333, 83.33333333, 73.33333333, 83.33333333, 50, 86.66666667, 
73.33333333, 86.66666667, 79.31034483, 70, 70, 66.66666667, 60, 
66.66666667, 56.66666667, 73.33333333, 63.33333333, 70, 55.17241379, 
80, 93.33333333, 83.33333333, 73.33333333, 83.33333333, 50, 86.66666667, 
73.33333333, 86.66666667, 79.31034483), auc = c(0.761904762, 
0.888888889, 0.753968254, 0.7, 0.8, 0.4, 0.85, 0.675, 0.85, 0.697222222, 
0.658730159, 0.563492063, 0.603174603, 0.525, 0.6, 0.475, 0.675, 
0.55, 0.65, 0.461111111, 0.761904762, 0.888888889, 0.753968254, 
0.7, 0.8, 0.4, 0.85, 0.675, 0.85, 0.697222222)), row.names = c(NA, 
30L), class = "data.frame")

Upvotes: 1

Views: 461

Answers (1)

Martin Gal
Martin Gal

Reputation: 16988

I looked into the ScottKnott package and the SK help. The model argument is not described. Your code should run with

library(ScottKnott)

sk1 <- with(Data, 
            SK(kappa ~ classifier,
               which='classifier')
            )

returning

#> Results
#> Means G1 G2
#> U-KNNIG4  0.49  a   
#> U-KNNCFS  0.49  a   
#> U-KNNCON  0.16     b
#> 
#> Sig.level
#> 0.05
#> 
#> Statistics
#> lambda chisq dfchisq pvalue evmean dferror
#> Clus 1     13   7.2     2.6 0.0038 0.0063      27
#> Clus 2      0   5.5     1.8 1.0000 0.0063      27
#> 
#> 
#> Clusters
#> #################  Cluster  1  ################
#> {G1}: U-KNNIG4 U-KNNCFS
#> {G2}: U-KNNCON
#> #################  Cluster  2  ################
#> {G1}: 0
#> {G2}: NA

Upvotes: 1

Related Questions