The data reader function for MAMS datasets are added in cmn/mams.py. The function is tested against lda for the test.xml file and runs without any syntax or run-time errors. The log file and final evaluations' means in different folds are attached below:
The MAMS dataset (ATSA) has some "error" in the annotation for the aspects. For example, an aspect term may be listed as "beer s" instead of "beer" or "beer selection". As such, an unaltered version of the datasets are kept in the data folder for any future reference while the edited version is currently in use.
hosseinfani
commented
1 year ago
What is accomplished?
The data reader function for MAMS datasets are added in cmn/mams.py. The function is tested against lda for the test.xml file and runs without any syntax or run-time errors. The log file and final evaluations' means in different folds are attached below:
test-run-log.txt pred.eval.mean.csv
What should be noted?
The MAMS dataset (ATSA) has some "error" in the annotation for the aspects. For example, an aspect term may be listed as "beer s" instead of "beer" or "beer selection". As such, an unaltered version of the datasets are kept in the data folder for any future reference while the edited version is currently in use.