<![CDATA[The study aims to evaluate the feelings of ChatGPT users based on the given rating, using lexicon methods to classify feelings in user comments without requiring prior class labels. This lexicon method uses the vader lexicon dictionary to calculate the sentimental polarity of words in comments. The comment data set is obtained from an external source, namely the Kaggle data set, which did not have a previous sentiment label. The lexicon model is applied as an unsupervised learning approach to understanding the expression of user sentiment. The comments in the dataset are processed and analyzed using a lexicon model. The results of this sentimental analysis show that the lexicon method produces comments that tend to be positive or negative for the rating range one to five. Nonetheless, research findings suggest that some aspects of sentiment may not be accurately detected by this method. Nevertheless, this sentiment analysis provides an important basis for further understanding of the interaction between users and applications. These results can be the basis for more in-depth research involving more sophisticated approaches to sentiment analysis.]]>
Copyrights © 2024
