- Developed a hybrid model combining 3D-CNN and LSTM networks to predict snow with 97.7% accuracy
- Designed a multimodal architecture processing both - radar images and numerical meteorological data for snow prediction
- Created a novel data preparation function to transform the numerical data into sliding window of observations to predict snow 3 days in advance based on 5 days of historical data
- Boosted the overall accuracy of Lake-effect snow prediction by 43%, surpassing the baseline decision tree model
- Zeiler. et.al 2013, "Visualizing and Understanding Convolutional Networks" (https://arxiv.org/abs/1311.2901)
- https://medium.com/@mixanyy/different-ways-to-combine-cnn-and-lstm-networks-for-time-series-classification-tasks-b03fc37e91b6