为什么我在WEKA中读取该ARFF文件会导致EOF早熟？

Question

Here's a reduced version I am using for testing that produces the error below

@RELATION documents

@ATTRIBUTE rnClass {r,n}
@ATTRIBUTE CONTAINS_EMAIL numeric
@ATTRIBUTE CONTAINS_PHONE numeric

@DATA
n,0.0,0.0
n,0.0,0.0
r,0.0,0.0
r,0.0,0.0
n,0.0,0.0
r,0.0,0.0
r,0.0,0.0


Caused by: java.io.IOException: premature end of file, read Token[EOF], line 1
    at weka.core.converters.ArffLoader$ArffReader.errorMessage(ArffLoader.java:258) ~[weka-dev-3.7.6.jar:na]
    at weka.core.converters.ArffLoader$ArffReader.readHeader(ArffLoader.java:678) ~[weka-dev-3.7.6.jar:na]
    at weka.core.converters.ArffLoader$ArffReader.<init>(ArffLoader.java:139) ~[weka-dev-3.7.6.jar:na]

However this previous version seems to be working without issues:

@RELATION documents

@ATTRIBUTE rnClass {r,n}
@ATTRIBUTE text String

@DATA
r,'blah blah blah...'
r,'\n\n\n\n\t\n\nblah blah blah\n\n...'
n,'\n\n   \n\n\n \nblah blah blah \n\n\nblah blah \n \n\n\n \'

Answer 1

I tried creating an arff file with your input, and its working for me. It may be the case that you arff file is corrupted. Can you create a new arff file and try ?

Answer 2

I had exactly this problem when trying to generate some files using the Notepad++ to make a data mining experiment. I suspected that could be an encoding problem that somehow corrupts the file and makes WEKA stop working. You should try to change the file encoding configuration: if you create the file with UTF- 8 encoding, you should change to ANSI or another one. It worked.