I worked as a Research Associate Data Scientist in Professor Nicholas Tatonetti’s team at the Cedars-Sinai Center during the summer of 2023, as part of my second year at ENSTA Paris. The Tatonetti Lab (TLab) focuses on enhancing drug safety through advanced data analysis, with the aim of identifying and preventing adverse drug reactions. You can learn more about their groundbreaking work on their website.
Over the course of three months, I contributed to key projects aimed at improving drug safety. My primary contribution involved developing a model to detect novel drug-drug interactions by analyzing extensive databases of reported adverse events, supporting the lab’s mission to make medications safer for patients.
This repository contains some of the codes I used during my research internship at Cedars-Sinai.
The Report.pdf is my research report submitted to the university. It provides a detailed description of the entire project, covering all aspects of my work.
The full_training.ipynb notebook outlines the algorithm for the complete training process of models, including hyperparameter optimization. This script enables the training of models on a large number of adverse events.
The Training for corroboration.ipynb notebook is designed to train specific models and obtain their corresponding outputs, which are then used to carry out validation on those models.
Finally, the ttest.ipynb notebook contains the code I used to perform the t-test as part of the model validation protocol.
- Machine Learning Programming: Developed and implemented machine learning algorithms for data analysis.
- Large Database Processing: Gained experience in processing and analyzing large datasets of adverse event reports.
- Data Analysis: Conducted thorough analyses to uncover novel drug-drug interactions.
- Model Development: Built and optimized models for predicting outcomes based on historical data.
- Statistical Testing: Performed t-tests and other statistical analyses to validate model performance.
- Hyperparameter Optimization: Implemented techniques to fine-tune model parameters for better accuracy.
- Collaboration and Communication: Worked closely with a research team and effectively communicated findings.