Sentiment Analysis, or Sentiment Classification, is valuable for answering lots of different research questions, such as: How does public sentiment react to different tragic events? Do tragic events always elicit the same negative emotional states? What publicly expressed emotions help politicians get elected? Have the most common emotions changed over the last 150 years?
In this 1-hour workshop, we will use TDM Studio to learn about Sentiment Classification and also run two approaches to the task using Python in Jupyter Notebooks. One approach will be a baseline dictionary-based approach and the other a more recent SBERT-based model that performs at the state of the art for newspaper sentiment classification.
- Login to TDM Studio
- Create a dataset related to your topic of interest
- Run Dictionary-Based Sentiment Analysis
- Run and compare with SBERT-based model
This workshop is limited to UC San Diego affiliates. Please register with your @ucsd.edu email address.