Flow Prediction
Predict The flow of OD matrix using Gravity Model (GM) and Neural Network (NN)
Related Fields: Transportation Engineering, Traffic Engineering, Machine Learning, Neural Network with Pytorch
Project Overview
Alright, buckle up! 🚀 In this adventure, we’re diving into a fascinating comparison of the Gravity Model (GM) with some cutting-edge tech: Neural Networks (NN) and Graph Neural Networks (GNN). Picture this as a quest to uncover the hidden superpowers and kryptonites of each model, and maybe, just maybe, discover where they shine the brightest.
We’ve got our hands on some juicy datasets from Sioux Falls, Anaheim, and Chicago. These aren’t just any datasets; they’re the secret maps to our treasure hunt, packed with real-world urban magic. 🌆 With these treasures, we’re ready to explore how GM, NN, and GNN stack up against each other in the bustling streets of different cities. By delving into this data, we’re not just crunching numbers – we’re looking for those hidden gems of insight, those aha moments that reveal the mysteries of urban life through the lens of these models.
So, let’s get ready to draw some bold conclusions, spot intriguing patterns, and maybe even challenge what we thought we knew about these urban landscapes. Who knew data analysis could be this exciting, right? 😄
Project Steps :
- Code the GM model. We did this with 4 different approaches: 1-exponential 2-power 3-tanner 4-guess
- The last approach (Guess) is made by me and is a new approach.
- Model the NN and GNN
- connect all details to get the results all together.
Results of The Project
We are still working on the conclusion.
Code Sources :
This model was built with PyTorch.
Here you can find all the code and results :
Last Changes: