majogamit / Academic-Situation-Sentiment-Analysis

A Natural Language Processing reseach that aims to analyze sentiments of approximately 15000 Tagalog tweets regarding Academic situations (online, hybrid, face-to-face)using Keras library.

0 stars 0 forks source link

Label Tweet Sentiments #1

Closed majogamit closed 1 year ago

majogamit commented 1 year ago

Tagalog words cannot be processed by libraries of sentiment analyzer. To solve it, do the ff:

[x] Try using textblob
[x] Create new column for sentiments
[x] Use filipino badwords corpus from github to label negative sentiments
[x] Label one by one [NOT POSSIBLE]
[x] Create new code to compute polarity using pre existing tagalog lexicon

majogamit commented 1 year ago

Textblob outputs more positive than nltk sentiment analyzer

majogamit commented 1 year ago

Badwords return around 1000 negative sentiments only.

majogamit commented 1 year ago

This issue is solved but results are not promising. lmao