Improving E-Commerce Conversion Rates with Machine Learning
We employed machine learning models such as logistic regression, decision tree, random forest, and XGBoost to understand and predict conversion rates across different user and advertiser segments on an e-commerce site. Based on the prediction results, we provide actionable recommendations help business optimize their marketing strategies and improve user conversion rates.
Model Selection for Infectious Disease Dynamics
We tested the SINDy model selection algorithm on time-series data from deterministic and stochastic infectious disease models, and used SINDy to construct dynamical systems models with early experimental data from the reported COVID-19 cases in Hubei province, China.
Causality Detection using Cross Convergent Mapping
We tested causality for time-series data from deterministic systems using Cross Convergent Mapping (CCM) and conducted sensitivity analysis to evaluate the impact of time delays and “shadow” manifold dimensions on CCM performance.
Mathematical Modeling of Traffic Jams in Houston
We assessed the vehicular network in Houston, Texas, using data provided by Française des Jeux Mathématiques and Société de Calcul Mathématique SA. The Greenshields and Lighthill-Whitham-Richards (LWR) models were implemented to analyze Houston traffic flows.
MSc Research: Turing Patterns in Spatially-Extended Population Models
Turing and Turing-Hopf patterns in spatially extended Rosenzweig-MacArthur predator-prey models with allochthonous resources (https://doi.org/10.1007/s11538-019-00667-0).
Mathematical Modeling of Wound Healing
Mathematical model for epithelial spreading to predict the deformation of epithelial tissues, the concentration of active chemical species, and how they interact with other.