Projects

I have done some interesting projects in my spare time. Some of them are my personal interest and some of them are related to my research projects.

• Propensity Score Matching for Fake News The implementation repository for my paper Improving Generalizability of Fake News Detection Methods using Propensity Score Matching. For more details, please kindly refer to the README file in the repository page.

• Visual Poet: This is the poetry recommendation system that we built for the paper Through The Eyes of A Poet: Classical Poetry Recommendation with Visual Input on Social Media. You can upload a photo to get a poetry recommendation. For my japanese class project, I changed the Chinese classical poetry in the database to Japanese Haiku that I scraped from online sources. The app is built with Flask and deployed with Google App Engine.

• Log-Normal Mixture Implementation: For our KDD20 project, we want to model the distribution of normal users and mallicious users in Tencent Weibo data. Scikit-learn only provides mixture model for gaussian distribution. For this purpose, I implemented log-normal mixture model with Expectaion-Maximization algorithm.