During the pandemic, many activities that make people feel happy are no longer viable due to the
public health and crisis control regulations. How to recommend suitable alternative activities to people would be a challenge. As different happiness activities would have different implicit conditions, making
recommendation directly based on sematic similarity might not be practical for people. This project aims to explore new methods to encode relations and similarities between activities to enable more flexible, personalizable recommendations.
- Analysis and do text mining on happiness moments from HappyDB
- Try and test existing commence knowledge base tools
- Use python for data analysis
- Get familiar with commence knowledge analysis in natural language processing (NLP)