-
Notifications
You must be signed in to change notification settings - Fork 13
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Text experiements: test sentiments #465
Comments
Blocking Text experiments dianna-ai/dianna-exploration#175, dianna-ai/dianna-exploration#184, dianna-ai/dianna-exploration#183 and dianna-ai/dianna-exploration#182 |
Related to dianna-exploration PR 159. @WillemSpek don't forget to link the issues here with the PRs in the other repo ;-) |
Also, there's something odd about this PR (159) - it links to the images code, not to the text one. I cannot locate the work you did for the text (it's not the the Text branch). Please, fix the code <-> PR link. |
Compiled a list of words (adjectives) from the Stanford movie reviews dataset to chose test data from for the Lorentz workshop ICT with industry usecase. Most of the words appear witht he same score over the reviews int he dataset! When I have found different score that's indicated next to the word. list of sentiment adjectives found in the Stanford move reviews sentiment dataset and modelword positivity score (scale is from 1 (max negative) to 25 (max positive))word positivity score(s) baaaaaad 1 disgusting 1.75 irritating 2 vulgar 2.3 dissapointing 3 4.75 (combined with sligthly?) pointless 3.3 depressing 4.75 dull 5 boring 5.3 stupid 5.75 monotonous 6 terrible 6.3 unimaginative 6.75 nasty 7 pitiful 7.3 awkward 7.75 clunky 8 rotten 8.3 shrewd 8.75 13.3 ugly 9 cliched 9.3 pretentious 9.75 overwhelming 10 obvious 10.3 bewildered 10.75 awful 11 coarse 11.75 12.3 dark 11.75 freak 11.3 restrained 12 spiritless 12.3 conventional 12.75 serious 13 melodramatic 13.3 earnest 13.75 fast 14.3 ballistic 14.75 driven 15 smooth 15.3 artful 15.75 silly 16 cerebral 16.3 acclaimed 16.5 convinient 17 exceeds 17.3 curious 17.75 subtle 18 poetic 18.3 good 18.75 cinematic 14.3 19 fun 19.3 good 19.75 clever 20.3 engaging 20.75 better 21 (also 19.75) pretty 22 dazzling 22.75 great 23.5 fabulous 24 brilliant 24.3 perfection 24.75 masterpiece 25 |
Simplified to integer values and adjectives only: word score worthless 1 |
Design a list of at least 25 positive and negative words using the sentiment scale of the Stanford sentiment treebank.
One way to find out the sentiments is to use the browsing capability in the dataset above by limiting the sentence length to 3 words (possibly more tokens).
Another might be to look at the indexed original dataset's sentiment labels (normalized between 0 1nd 1)-
Here is the parent link to the original and derived datasets on Surfdrive.
Create 25 test sentences of length 3 containing the above words, one sentence each. Perhaps as simple as
Store the words in this issue.
Store the sentences in a .tsv file in the same format as the model's test data on Surfdrive.
Stems from #445 and dianna-ai/dianna-exploration#187 (see for Practicalities).
The text was updated successfully, but these errors were encountered: