Erdos Institute Data-Science Projects

During my work in the data science bootcamps at the Erdos Institute, I worked on two projects with a team of very cool collaborators in projects in data science in machine learning.

In Fall 2022, I worked on machine learning prediction models to predict copayment information based on patient history. When a patient is prescribed medication from healthcare providers, their net copayment at the end of the transaction is determined by a complex system involving specific drug treatments, insurance, and other pharmaceutical factors. Currently, patients and doctors do not have a method of checking expected costs before prescribing medication. We used various machine learning techniques including random forests and gradient boosting to predict copayments on synthetic datasets with accuracies of 90%. You can see our presentation, code on github, and writeup.

In Summer 2022, I worked on clustering and classification models for studying the What’s Cooking dataset from Kaggle in order to predict the cuisine of a dish based on its ingredients as well as clustering similar cuisines with various clustering techniques. The aim of the project was to study this dataset in order to better inform food and restaurant recommendation systems and we achieved classification accuracies of 80% with clustering visualizations as well. You can see our presentation, code on github, and writeup.

Karan Srivastava
Karan Srivastava
PhD Student, Mathematics

My research interests include machine learning, reinforcement learning, combinatorics, and algebraic geometry