The goal of this thesis is to analyze the performance of different machine learning and deep learning models in predicting sentiment related to hotel and b&b reviews near Pisa (Italy). The models tested were the following: SVM, Random Forest, Logistic Regression, CNN, LSTM.
| MODEL | PRECISION | RECALL | F-SCORE |
|---|---|---|---|
| SVM | 0.92 | 0.91 | 0.92 |
| RANDOM FOREST | 0.70 | 0.73 | 0.71 |
| LOGISTIC REGRESSION | 0.76 | 0.73 | 0.73 |
| CNN | 0.82 | 0.82 | 0.82 |
| LSTM | 0.80 | 0.79 | 0.79 |
At the end of the project, it was shown that the use of more complex models does not always guarantee better performance than the use of classical predictive models. In this specific case, the SVM model performed 15% better than the results obtained by applying neural networks.