In panel data forecasting, no single model consistently performs well across diverse panel data characteristics, such as varying entity counts, time steps, and data distributions. Traditional models often struggle to adapt to non-stationarity, diverse patterns, and complex entity interactions within panel data. This limitation underscores the need for robust deep-learning solutions that can handle these challenges and improve forecasting in critical fields like economics, environmental, transport and healthcare. Therefore, developing generalizable, robust deep learning architectures that can adapt to varying panel data characteristics is crucial to improve forecasting accuracy in these critical fields.
In our project, we are going to create a deep-learning model that performs well in panel datasets with different characteristics.
Kajaani Balabavan
A fourth-year undergraduate at the Department of Computer Science and Engineering at the University of Moratuwa. I am passionate about Data Science, with a particular interest in Machine Learning and Deep Learning for time series analysis, and I am currently exploring Panel Time Series Analysis. My engineering background and interest in data analysis drive me to develop innovative solutions for extracting insights from complex data.
Brief Description of Project
Recent News
- All
