Options
special attributes afetr the application of a classification model
ilaria_gori
Member Posts: 15 
Maven
Hi all,
I have a problem with the special attributes afetr the application of a classification model:
I remarked that, if I have non-numeric special attributes (for example string) in my data and I apply a classification model the resulting file with output has not all the special attributes (ex. the id of a lesion) written in the correct way.
Could you explain me why? It's important for me in order to make statistics (example: which lesion has been well classified etc..) on the result file.
Moreover I remarked the same kind of problem when I load two different file with label ( 1 for lesion and 0 for nonlesion) for tuning and prediction: if the first file has lesion examples before nonlesion examples and the second file has the contrary, in the file with output the label is not correct.
thank you very much for your help!
ilaria
I have a problem with the special attributes afetr the application of a classification model:
I remarked that, if I have non-numeric special attributes (for example string) in my data and I apply a classification model the resulting file with output has not all the special attributes (ex. the id of a lesion) written in the correct way.
Could you explain me why? It's important for me in order to make statistics (example: which lesion has been well classified etc..) on the result file.
Moreover I remarked the same kind of problem when I load two different file with label ( 1 for lesion and 0 for nonlesion) for tuning and prediction: if the first file has lesion examples before nonlesion examples and the second file has the contrary, in the file with output the label is not correct.
thank you very much for your help!
ilaria
0
Answers
This problem is well known, please see this thread here:
http://rapid-i.com/rapidforum/index.php/topic,776.msg2896.html
hope this helps
kind regards,
Steffen
PS: Another example where a wiki would be quite useful ;D