How do I fix NARX neural network performance value?

I am using NARX neural network to forecast wind speed. i am obtaining performance but my performance value changed when i retrained my model. i retrained my model many time but performance value little change. how can i solve this issue?

Thank you

Answer 1

[Hornett]

The performance value of your NARX neural network changes upon retraining due to the following reasons:

1. Random Initialization: Neural networks typically initialize weights randomly, leading to different starting points for each training run.
2. Stochastic Training Process: Algorithms like stochastic gradient descent introduce randomness in the training process.

Solutions to Achieve Consistent Performance

1. Set a Fixed Random Seed: Ensures reproducibility by initializing the random number generator to a fixed state.
2. Increase Training Epochs: Allows the network more time to converge, reducing variability.
3. Cross-Validation: Provides a more reliable performance assessment by averaging results over multiple data splits.
4. Ensemble Methods: Training multiple models and averaging their predictions can stabilize performance.