To accurately classify tweets related to ChatGPT as either positive or negative sentiments, helps us understand the public perception better.
In this sentiment analysis project, my goal was to accurately classify tweets as positive or negative. I began by conducting Exploratory Data Analysis (EDA) to gain insights into the textual data.To enable machine and deep learning, I converted the text into vector representations. I used logistic regression as our baseline model for this binary classification task, with a primary focus on achieving high precision to minimize false positives. Additionally, I make use of word embeddings to capture context and built complex neural networks, including RNN, LSTM. Ultimately, the Bi-directional LSTM emerged as the top-performing model, achieving an impressive precision score of 0.93 for the test dataset.