This repo acts as a catalog and storage for my personal and course project reports. Below is a table of contents each with a brief overview of the reports. This table of contents will incrementally become larger as I organize the reports and files.
| Project Name | Keywords | Overview | Comments & Critiques | Start | End | Grade Received | Collaborator |
|---|---|---|---|---|---|---|---|
| Predicting Item Rating by Text Mining and NLP Analysis with Renttherunway Clothing Fit Dataset | Data Mining, Recommender System, Machine Learning, Natural Language Processing | Through this report, we summarized and presented our approach to develop a predictive recommender system iteratively, starting from the very baseline of a using a SVD and logistic regression model, up to adding word embedding using Continuous Bag of Words (CBOW) and SkipGram. At each step, we reasoned our intuition as to why we are building our parameters and models the way we did. In the end, we evaluated each model holistically using different error metrics and selected the one that performed best overall. We did not directly discuss the incorporation of our model into a functional recommendation system but it is our immediate next step of future works. | "They have explored a decent number of papers and explained how well their implementations are." | 9/2022 | 12/2022 | A+ | Yunfan Long |
| Who Will Get A Stroke? | Imbalance Data, Logistic Regression, Classification | Through this project, we aim to build a classifier that can predict the onset of stroke with high accuracy (hopefully near 90%). The objective may seem very intuitive and straightforward, but we want to highlight that we trained this classifier through a procedure called step-wise model selection, which we implemented from scratch. We iteratively train and select the best feature that boosts our model's performance the most. Ultimately, not only do we want to train a powerful classifier, we hope the features which constitute our most accurate classifier can serve as precautionary factors in preventing stroke with some grain of truth. | N/A | 3/2021 | 6/2021 | A | Wenqian Zhao, Cecilia Liu |