- Weather prediction using LSTM and Global Attention.
- Multivariate time series analysis; the dataset contains 14 different features such as atmospheric pressure, humidity, temperature, etc. The goal is to predict the future temperature using all these 14 features.
Steps to use the dataset.
- Click this link to download the dataset.
- Extract the zip file.
- Place the contents inside the extracted folder into the main repository.
After preparing the dataset, train the neural network:
python main.py
- Some sampled predictions from the test set after the model was fully trained.
- Here, the x-axis is the time in days, and the y-axis denotes the temperature in degC.