Projects

VR Sickness Prediction Based on Time-Series Forecasting

To enhance the VR user experience, it is important to predict VR sickness before it occurs and develop techniques to mitigate it. In this project, I developed VR sickness prediction models using head-mounted display (HMD) sensor data, such as eye movements, head movements, and video optical flow, based on time-series forecasting and LLMs.

Development and Data Analysis of VR-Based DTx for Depression

We developed a VR-based digital therapeutics (DTx) for depression based on clinical evidence (i.e., acceptance and commitment therapy (ACT)), and evaluated its clinical feasibility. Through a user study, we objectively demonstrated the clinical effectivenss of this system by analyzing sensor data from VR and wearable devices.

Stock Price Prediction Using Time-Series Forecasting Models

This graduation project used S&P stock price data to build time-series forecasting models, and compare their predictive performance. Through this project, I deepened my understanding of time-series forecasting models, which later served as a foundation for sensor-based time-series prediction research during master's program.

Development of a Laundry Method Search Engine

In this team project, we developed a website that laundry shop owners can search for laundry methods shared within their community. I built a database and designed its schema using data from social network (i.e., BAND) posts, and designed an intuitive UI considering the age group and usability needs of the main users.

[CJ Logistics] Development of a Parcel Delivery Demand Forecasting Model

To more precisely predict parcel delivery demand, I developed time-series forecasting and tabular deep learning models that incorporate external contextual factors, improving performance compared to the existing in-house model. I also implemented an AI agent-based XAI demo to help non-AI experts understand the prediction results.