search

SverreNystad / power-predictor

Using Machine Learning for time series forecasting of photovoltaic measurement for solar systems based on weather features
Other
6 stars 2 forks source link

LSTM (Long Short-Term Memory) #10

Closed SverreNystad closed 7 months ago

SverreNystad commented 9 months ago

Rationale: Given its capacity to remember patterns over long sequences, it might be especially useful if our data has temporal sequences or time-series components. Implementation: Consider using TensorFlow, pytorch or Keras for implementation. Potential Challenges: LSTMs can be computationally intensive and might require more time to train.

SverreNystad commented 9 months ago

Notebook with impl: https://colab.research.google.com/drive/1b3CUJuDOmPmNdZFH3LQDmt5F0K3FZhqD?usp=sharing

SverreNystad commented 9 months ago

We need to experiment with different layer design for this model. Right now we only have Features into a Dense layer of 1 to predict pv_measurement

SverreNystad commented 9 months ago

The Window size determines how far into the distance we want the model to predict. Need to find what would be an ideal amount.

SverreNystad commented 9 months ago

For overfitting use: early stopping, dropout, or regularization

SverreNystad commented 9 months ago

Try not to use it to get exact pv measurements but rather find the difference between different pv measurements