Good afternoon! Having trouble with t+1 price estimation using Hidden Markov Models, this article is the guide https://rubikscode.net/2018/10/29/stock-price-prediction-using-hidden-markov-model/. It takes a sample from training data, adds a possible t+1 outcome, and scores the model with hmmlearn.hmm.score. The scores vary but the location of the max (np.argmax) is always at the same location, regardless of the data sample. I know it is a pretty vague question, but I figured if anyone could find the error it would be you all! Thank you in advance for any guidance!
Submitted February 26, 2020 at 02:26PM by conhardy