This is the repository for the thermal expansion error prediction research project done during Summer 2023. This project uses the LSTM network for training and regression. Most of the programming and tidying work were done through Python Keras & Scikit-learn & Pandas, which can be found in the lstm_keras directory. Some other MATLAB implementations can be found in the lstm_matlab directory. For more details, please refer to the report. Please feel free to email me at ruoheng.du@nyu.edu for any more information.
The best model done by LSTM network has a train RMSE of 0.126 and a test RMSE of 0.127. More details can be found under the /lstm_keras/best model/ directory.
