A carregar...
Publicação
Using Consumer Online Reviews to Predict Stock Price Returns in the Ridesharing Industry
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 2.03 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
This research explores the predictive power of online consumer reviews on weekly stock price
returns within the ridesharing industry, focusing on Uber and Grab, two companies with very
distinct characteristics. The user-generated reviews were obtained from the Google Play Store
and leveraged to determine how their features (such as sentiment, rating, and textual
content) correlate with stock performance. This data is joined with Stock data from Yahoo
Finance and aggregated daily and weekly to be inputted into machine learning models,
specifically Recurrent Neural Networks with Long Short-Term Memory (RNN-LSTM) and
XGBoost. After identifying the optimal time lag for the predictions, the results reveal high
predictive capacity from some sets of features, mainly regarding the review content itself.
These findings confirm the benefits of considering user feedback in financial forecasting by
showing that including features such as the most common words and the rating given by users
increases performance for stock price prediction, especially in terms of the price return
direction (positive or negative). By demonstrating the value added in integrating qualitative
online review data into quantitative stock prediction, this work contributes to financial
analytics, offering relevant insights for investors and industry stakeholders to understand how
their customers influence their economic outcomes.
Descrição
Dissertation presented as the partial requirement for obtaining a Master's degree in Data Science and Advanced Analytics, specialization in Data Science
Palavras-chave
User Generated Content Stock Price Returns Prediction Deep Learning Ridesharing SDG 8 - Decent work and economic growth SDG 9 - Industry, innovation and infrastructure