I’ve been working on this stock price prediction model. There’s an issue that I am facing while trying to fit the model. Everytime I try to train the model, the loss function shows a sharp decrease in the first epoch but remains stagnant in the following epochs. I’ve tried switching up the size of the data, the number of features engineered, scaling and not scaling the features/target, switching up the hyperparameters, nothing seems to work. The predictions have a very small range (usually around the mean) compared to the actual target values’ range. Any help is appreciated.
The first place to look is to verify that in your model, that is is correctly reading in the training data, and is using the expected labels.