Hello,
When I load a dataset with 2 string attributes using weka's Explorer, and build a classifier using SMO/StringKernel, it seems that weka automatically merges the string attributes into one.
But when I load the same date with the java api, it only use the first attribute and ignore the other one...
Is there a way I can reproduce this behavior using the java API?
When I load a dataset with 2 string attributes using weka's Explorer, and build a classifier using SMO/StringKernel, it seems that weka automatically merges the string attributes into one.
But when I load the same date with the java api, it only use the first attribute and ignore the other one...
Is there a way I can reproduce this behavior using the java API?
