Predict hotel review sentiment and predict star ratings using fine-tuned RoBERTa with ordinal classification and embedding-based analysis.

Transformer-based modeling of hotel review text to predict star ratings (1–5), evaluate ordinal classification performance, and analyze sentiment separability through embedding-based clustering.

A reproducible LSTM forecasting implementation for the log-transformed Sentiment–Volatility Ratio derived from UMCSI and VIX, featuring data preprocessing, time-series modeling, and empirical evaluation of sentiment-driven market volatility dynamics.

Standalone AR forecasting workflow for the Sentiment–Volatility Ratio capstone project, using UMCSI and VIX data with expanding-window evaluation.

Standalone repository for preserving, validating, and modernizing the multilayer perceptron forecasting workflow used to predict the log-transformed Sentiment–Volatility Ratio at one- and three-month horizons.