ajinkyat / deep_learning_entity_embeddings

Deep Learning and Entity Embeddings to predict driving behaviour and cluster accident hotspots

Deep Learning and Entity Embeddings to predict driving behavior and cluster accident hotspots

This project repo is for Pune Smart City Hackathon organized by Niti Aayog (Planning Commision of India), Government of India

Check below Jupyter notebooks for solution overview.

Data cleaning and processing: data_processing.ipynb

Model architecture: model_building.ipynb

Visualizing trained embeddings: embedding_visualization.ipynb (Plotly visualizations are not rendered properly) instead download embedding_visualization.html and open it locally in the browser

Visualizing Route-Name Embeddings: https://plot.ly/~ajinkyaT/5/

Visualizing Stop-Name Embeddings: https://plot.ly/~ajinkyaT/3/

Resources

• Guo, C., & Berkhahn, F. (2016). Entity embeddings of categorical variables. arXiv preprint arXiv:1604.06737
• De Brébisson, A., Simon, É., Auvolat, A., Vincent, P., & Bengio, Y. (2015). Artificial neural networks applied to taxi destination prediction. arXiv preprint arXiv:1508.00021.