PROJECTS
CMS ARTIFICIALÂ INTELLIGENCE HEALTH OUTCOMES CHALLENGE
"Multi-Layered Feature Selection and Dynamic Personalized Scoring"
Participated as a member of the data science team, under the leadership of Dr. Shashaani and Dr. Swann. We combined an integer programming based risk scoring algorithm with Machine Learning clustering and Simulation Optimization based feature selection to predict unplanned hospital admission for 30-days intervals.
Our team was one of top 25 ideas selected for the stage 1 of the
competition. We published and presented part of our work in the 2020 INFORMS Service Science conference.
CALL FOR CODE SPOT CHALLENGE: MENTAL HEALTH IN A TIME OF CRISIS
"Data-driven Recommendations for Equitable Mental Health Resource Allocation"
Worked with a group of PhD students at the ISE and OR departments. Our team was selected as one of the top 5 finalists. This challenge was hosted by IBM collaborating with Anthem, Beacon Health Options, and XPRIZE.Â
We developed a management dashboard for the state-level decision makers to allocate monetary resources to different counties, focusing on the minority population in each region.
ACADEMIC POSTERS
"Improved Feature Selection with Simulation Optimization"
I have presented my research on improved feature selection via simulation optimization in many conferences such as, INFORMSÂ 2020, WSC 2019 and WSC 2020. My poster has been selected as one of the finalists in the poster competition of INFORMS 2020 as well.Â
Feature selection is the task of identifying uninformative and redundant variables in a dataset. In this work, we attempt to conduct more accurate loss function evaluation to improve upon well-known heuristic optimization algorithms in feature selection.
DATA SCIENCE IN INDUSTRIAL ENGINEERINGÂ COURSE PROJECT
"AD Big Data DREAM Challenge"
In a team of 4 PhD students, we used the "AD Big Data DREAM Challenge" data to implement various Machine Learning models such as KNN (K-nearest neighbors), SVM (support vector machines) and generalized linear regression.
Employed R to implement the mentioned models and gained close accuracy rates to the winners' of this competition.
LOGISTIC ENGINEERING COURSE PROJECT
"Optimizing Routes for NCSU Wolf-line Buses"
Worked in a team of two on our final project which was to optimize the school bus lines, considering common students living areas and campus layout. Used MATLAB and Gurobi to employ online and offline optimization heuristics to construct routes for NCSU buses.
Compared our results with current routes and showed that our model can reduce the number of buses while maintaining the average waiting time.
BACHELORS SENIOR PROJECT
"Implementation of Forecasting Models in Financial Engineering"
Implemented different models such as GARCH, ARCH and ARIMA in forecasting Iran's Market stock's values, using SAS. Achieved more than 80% accuracy rate in 2 weeks prediction window.
