Negation with "not" and "no" is ignored during sentiment analysis
Maintainers usually reply within 2 days
Assessment
- Difficulty
- 2/5
- Estimated time
- 1-3 hours
- Newbie friendliness
- 68/100
- Issue type
- Bug
- Clarity
- Mostly clear
- Activity status
- Active
- Tech stack
- python
- Domain
- machine-learning
Research direction
Read the README’s description of negation, then trace where stopwords are removed and where negation is handled during tokenization. Try the reported pairs (fresh/not fresh and stale food/no stale food); done means the negation words are not discarded before they can affect the sentiment result.
Written by the indexing model from the issue text.
Description
I noticed that sentences using not or no are being interpreted the same
way as sentences without the negation.
For example:
fresh
and
not fresh
both get tokenized as:
["fresh"]
and end up producing the same positive prediction.
The same thing happens with:
stale food
and
no stale food
According to the README, not, no, and never should negate the word that
comes after them, so I expected these inputs to affect the sentiment result.
It looks like not and no are included in the stopword list and get removed
before the negation logic gets a chance to process them. Because of this,
phrases like not delicious can be interpreted exactly like delicious,
which can result in the opposite sentiment from what the user intended.
- Dominant language
- Python
- Stars
- 0
- Forks
- 3
- Avg merge
- 1d 9h
- Merged PRs (30d)
- 13
Getting set up
- No Dockerfile or Docker Compose file
- Has a pull request template
- No contributing guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
More from techcsispit/mess-mood
-
Standalone punctuation breaks negation handling and reverses sentiment predictionsPossibly taken @vivek-3728 claimed this 1 day ago. Opengood first issue
Difficulty 2/5 1-3 hours Newbie friendliness 87/100
techcsispit/mess-mood#28 · 1 comment · 1 assignee ·
Maintainers usually reply within 2 days
-
k_fold: 3 of the 113 reviews are never in any 5-fold test foldPossibly taken @vivek-3728 claimed this 14 days ago. Openadvanced bug
Difficulty 2/5 1-3 hours Newbie friendliness 25/100
techcsispit/mess-mood#14 · 2 comments · 1 assignee ·
Maintainers usually reply within 2 days
All issues in techcsispit/mess-mood
Similar issues
-
bug
Difficulty 2/5 1-3 hours Newbie friendliness 78/100
Maintainers usually reply within 1 day
-
Difficulty 2/5 1-3 hours Newbie friendliness 72/100
aws/sagemaker-python-sdk#6554 ·
Maintainers usually reply within 1 day
-
Difficulty 1/5 Under an hour Newbie friendliness 78/100
huggingface/diffusers#15170 ·
Maintainers usually reply within 1 day
-
[BUG] LLM conflict check ignores MNEMOSYNE_LLM_MAX_TOKENS, so reasoning models return empty repliesOpen
Difficulty 2/5 1-3 hours Newbie friendliness 62/100
mnemosyne-oss/mnemosyne#1325 ·
Maintainers usually reply within 1 day
-
enhancement meshtastic-2.8
Difficulty 2/5 1-3 hours Newbie friendliness 62/100
pskillen/meshflow-api#459 ·