Neural Network-based Time Series Analysis of Ice Dataset

This work uses shallow neural networks to predict global ice volume from a 440,000-year dataset. By learning historical trends, the network.

This work explores shallow neural networks for time-series prediction, focusing on modeling global ice volume. Using the ice_dataset, containing 219 measurements of ice volume over 440,000 years, a neural network is trained to predict future ice volume based on past values. This single time-series prediction problem leverages the network's ability to learn patterns in historical data. By analyzing past ice volume trends, the network can forecast future changes, contributing valuable insights to climate studies and our understanding of Earth's glacial cycles.