search

Waikato / meka

Multi-label classifiers and evaluation procedures using the Weka machine learning framework.
http://waikato.github.io/meka/
GNU General Public License v3.0
200 stars 76 forks source link

Illegal options -t #74

Closed zhangky12 closed 3 years ago

zhangky12 commented 3 years ago

I tried to run the example code simply with

java -cp "./lib/*" meka.classifiers.multilabel.BR -t data/Music.arff

But got the following error message

java.lang.Exception: Illegal options: -t data/Music.arff

Help will be appreciated.

fracpete commented 3 years ago

Looks like recent changes in Weka screwed up the command-line use of Meka. The latest version of Weka enforces the check for remaining options when configuring a classifier now. In Meka's Evaluation class, the classifier got configured before accessing the general command-line options, resulting in this error.

I just committed a fix for that (Meka's Evaluation class now extracts all general options from the options array now before configuring the classifier).

Please download the latest snapshot and try again: https://adams.cms.waikato.ac.nz/snapshots/meka/

zhangky12 commented 3 years ago

Looks like recent changes in Weka screwed up the command-line use of Meka. The latest version of Weka enforces the check for remaining options when configuring a classifier now. In Meka's Evaluation class, the classifier got configured before accessing the general command-line options, resulting in this error.

I just committed a fix for that (Meka's Evaluation class now extracts all general options from the options array now before configuring the classifier).

Please download the latest snapshot and try again: https://adams.cms.waikato.ac.nz/snapshots/meka/

It works now. Thank you!