Sentiment Analysis and Performance Forecasting Based on Online Review Hotel Harper M. T. Haryono Jakarta

Abstrak

This thesis examines the sentiment of online reviews for Hotel Harper to identify key areas for service improvement, predict occupancy rates using sentiment analysis and forecasting models, and provide strategic recommendations for enhancing the hotel's image and performance. The research involves literature review, data collection, sentiment analysis, and forecasting model development. Data from online reviews of Hotel Harper M.T. Haryono from 2020-2023, spanning periods before, during, and after the COVID-19 pandemic, were collected. The Naive Bayes algorithm classifies sentiments into positive, negative, and neutral categories. Data visualization and classification performance evaluation are also conducted. Sentiment data is combined with occupancy rate data to develop an ARIMA forecasting model, evaluated using MAPE and RMSE. Results indicate that room quality and cleanliness significantly influence user evaluations, necessitating improvements in these areas. Negative reviews pointing to service-related issues suggest the need for enhanced staff training. Consequently, Hotel Harper M.T. Haryono should conduct regular training sessions for staff, especially those interacting with guests, to improve service quality. The occupancy rate predictions show an upward trend from 2020 to 2023 with low error rates, enabling Hotel Harper M.T. Haryono to use this model for strategic planning and informed decision-making.