search

huseinzol05 / Stock-Prediction-Models

Gathers machine learning and deep learning models for Stock forecasting including trading bots and simulations
Apache License 2.0
7.84k stars 2.78k forks source link

Accuracy Failure #5

Closed bitnom closed 5 years ago

bitnom commented 5 years ago

I got the script running by using Pipenv. Please see this readme and the Pipfile I made. The problem now is that I'm getting:

Best AGENT accuracy value: 3.244000

Is this because I am training on the GOOG.csv dataset? The Tesla file you used is not included in the repo. I'm wondering what the difference is and why I got low accuracy.

thanks

huseinzol05 commented 5 years ago

Herm, I am not really sure about this, I am going to commit this repository after I done committed with my company repositories ya

huseinzol05 commented 5 years ago

Oh, you trained on Google trend but predict on Tesla trend. Ya, you can't. the model doesn't have a good generalization between two different trends. It need to train real time, if trained on Google, it only can predict on Google.

Maybe you can try to train more than 10 different trends, and try to predict 11th trend, let see how is good the model generalization. Never tested this theory. Problem with reinforcement learning, let say the agent trained to play DOOM shooting game. After that you put the agent on different shooting game, counter strike. Even both are shooting game, but the agent never learn how to adapt inside counter strike environment. I never saw a reinforcement learning agent able to adapt different environment to test in any research papers, correct me if I am wrong, cited with paper link.

AlconDivino commented 5 years ago

I recommend using the GOOG-year.csv file as data. The GOOG.csv file has only 24 daily close prices i think its not strange to see failure in accuracy there

bitnom commented 5 years ago

Oh, you trained on Google trend but predict on Tesla trend. Ya, you can't. the model doesn't have a good generalization between two different trends. It need to train real time, if trained on Google, it only can predict on Google.

Maybe you can try to train more than 10 different trends, and try to predict 11th trend, let see how is good the model generalization. Never tested this theory. Problem with reinforcement learning, let say the agent trained to play DOOM shooting game. After that you put the agent on different shooting game, counter strike. Even both are shooting game, but the agent never learn how to adapt inside counter strike environment. I never saw a reinforcement learning agent able to adapt different environment to test in any research papers, correct me if I am wrong, cited with paper link.

I tried to pass them but how did I mess up the shape? https://github.com/huseinzol05/Stock-Prediction-Models/issues/7

bitnom commented 5 years ago

Oh, you trained on Google trend but predict on Tesla trend. Ya, you can't. the model doesn't have a good generalization between two different trends. It need to train real time, if trained on Google, it only can predict on Google.

Maybe you can try to train more than 10 different trends, and try to predict 11th trend, let see how is good the model generalization. Never tested this theory. Problem with reinforcement learning, let say the agent trained to play DOOM shooting game. After that you put the agent on different shooting game, counter strike. Even both are shooting game, but the agent never learn how to adapt inside counter strike environment. I never saw a reinforcement learning agent able to adapt different environment to test in any research papers, correct me if I am wrong, cited with paper link.

Finally got it working with the yearly and am passing the outputs automatically: https://github.com/TensorTom/Stock-Prediction-Models/blob/master/agent/evolution-strategy-bayesian-agent.ipynb