This repo is the final project of information retrieving. I tried Bayesian model and classic LSTM model to analyse the sentiment of reviews in both English and Chinese datasets.
The result of the two models can be found in file:report/info.ppt, and the source code can be found in file:src/.
Here is the accuracy of the 4 experiments:
| Bayesian model | classic LSTM model | |
| Chinese Dataset | 71.2% | 77.1% |
| English Dataset | 83.5% | 79.4% |