The ChnSentiCorp dataset of hotel reviews contains 9,600 reviews for training and 1,200 for testing. Each review is labeled either positive or negative.
- Develop a machine learning model to detect sentiment
- What can be done to improve the accuracy of the model you developed?
- Can you apply your model to another dataset and achieve similar accuracy?
Some relevant resources for sentiment analysis in Chinese could be found here.