Epidemiology with agent-based models Abstract
This project aims to present a computational simulation of disease transmission using agent based modeling techniques implemented in Python using the Mesa package. By developing a custom disease model, we investigate the effects of various factors such as infection rates, recovery types, age and vaccination availability on the dynamics of an infectious disease. The findings from our simulation aim to contribute to the broader understanding of disease transmission dynamics, which has been a major focus in various fields over the last several years.
Deep Learning Classification Models with AG News The growth of online news has led to a large volume of digital data. Natural language processing can help leverage this data for useful applications such as topic classification. Deep learning classification models are an attractive option to be used in these applications as the many layers in deep networks allow for hierarchal processing of the data (Glassner, 2021). The field of natural language processing has seen rapid growth since the advent of deep learning models with recurrent neural networks leading the charge (Ketkar & Moolayil, 2021).
Prediction of Bodyfat using a Linear Regression Model and Body Measurements Abstract
Bodyfat percentage is an important estimator for health. The most accurate method for measuring bodyfat is by underwater weighing which is time-intensive. Predictive analytics (linear regression) using indirect measurements offer a faster method to compute bodyfat percentage. This study analyzes bodyfat data determined by underwater weighing with their corresponding indirect measurements. To predict the bodyfat using these measurements, we compare a traditional linear model, regularized model, subset models (indicator, dichotomous, piecewise, polynomial), and feature engineering models (principal components analysis).
