SOCprediction

Initial Project for MARC

Use first 3 rows (voltage, current, temperature) to use as input for LSTM model to predict SOC

only use the first 100 000 data points

use 10 hidden units

1000 epochs

Notes:

-To create the model, I just copied general design from the MATLAB file.

-The predictions aren't as accurate when the SOC has a negative slope.

-The prediction is a bit different everytime I run it.

-Sometimes it will just be a horizontal line, if this happens I just run the code again and it normally works

-1000 epochs didn't seem to have much more of an effect compared to 100, I left it at 100 so its easier to debug but it can be easily changed

Name		Name	Last commit message	Last commit date
Latest commit History 85 Commits
.ipynb_checkpoints		.ipynb_checkpoints
Data		Data
Helper Programs		Helper Programs
ITEC_Comp		ITEC_Comp
Models		Models
Notebooks		Notebooks
Pictures		Pictures
Templates		Templates
.gitignore		.gitignore
README.md		README.md
SOC_LSTM_pytorch.ipynb		SOC_LSTM_pytorch.ipynb
SOCgoodcopy.ipynb		SOCgoodcopy.ipynb
SOCtest.ipynb		SOCtest.ipynb
autogluon_SOC.ipynb		autogluon_SOC.ipynb
autogluon_SOC_LSTM.ipynb		autogluon_SOC_LSTM.ipynb
autogluon_range.ipynb		autogluon_range.ipynb
main.py		main.py
pytorch_SOC.ipynb		pytorch_SOC.ipynb
pytorch_fnn_soc.ipynb		pytorch_fnn_soc.ipynb
requirements.txt		requirements.txt
routeapi.py		routeapi.py
server_autogluon_SOC.py		server_autogluon_SOC.py
server_autogluon_range.py		server_autogluon_range.py
xlsx_dataextract_SOC.py		xlsx_dataextract_SOC.py

Provide feedback