I'm starting off my undergraduate research on text classification even though I'm still considered new to this topic. I've collected more than 20K data from Twitter. I've been trying to label the data into 3 sentiments, positive, negative, and neutral. But, I oftentimes find it difficult to determine whether a tweet is categorized as positive, negative, or neutral due to my lack of knowledge in that field.

My supervisor has someone with a psychology background that can help me labeling the data instead of linguistics. I'm aware that sentiment analysis is a part of the natural language processing task which makes more sense if I ask for help from linguistics, but at the same time sentiment, in general, is also studied in psychology as I read more papers. Does anyone have any advice? Thank you in advance!


1 Answer 1


Psychology was mentioned because psychology has a long history of assigning numeric scores to subjective topics. One of the most important concepts is inter-rater reliability, how much do different people agree on an interpretation.

Other concepts that are useful are the degree of subjectivity and the degree of polarity (vs. assigning binary polarity labels). This is how Python's TextBlob package models sentiment.

  • $\begingroup$ Thank you for the answer, I appreciate it! 🙌 However, I've done my undergraduate research by involving both psychology and linguistic people at the same time. $\endgroup$
    – Faaiz
    Commented Jan 27, 2022 at 8:57

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.