revisit-severson-et-al
Repository for the paper "Statistical learning for accurate and interpretable battery lifetime prediction"
Science Score: 49.0%
This score indicates how likely this project is to be science-related based on various indicators:
-
○CITATION.cff file
-
✓codemeta.json file
Found codemeta.json file -
✓.zenodo.json file
Found .zenodo.json file -
✓DOI references
Found 6 DOI reference(s) in README -
✓Academic publication links
Links to: zenodo.org -
○Committers with academic emails
-
○Institutional organization owner
-
○JOSS paper metadata
-
○Scientific vocabulary similarity
Low similarity (12.1%) to scientific vocabulary
Keywords
Keywords from Contributors
Repository
Repository for the paper "Statistical learning for accurate and interpretable battery lifetime prediction"
Basic Info
Statistics
- Stars: 54
- Watchers: 3
- Forks: 15
- Open Issues: 2
- Releases: 4
Topics
Metadata Files
README.md
revisit-severson-et-al
NOTE: Please contact Prof. Richard Braatz, braatz@mit.edu, for access to the code repository associated with the Severson et al. publication in Nature Energy (available with an academic license). This repository is not directly related to the Nature Energy paper.
This repository contains code for our work entitled "Statistical learning for accurate and interpretable battery lifetime prediction", a follow-up paper to Severson et al. A permanent archive of this work on Zenodo is available here:
Our key scripts and functions are summarized here:
- generate_voltage_arrays.m: MATLAB script that generates capacity arrays from the battery dataset and exports them to csvs (stored in /data).
- revisit-severson-et-al.ipynb: Python notebook containing most of the analysis and figure generation.
- image_annotated_heatmap.py: Helper function from matplotlib (see docstring for source).
- lifetime_charging_time/lifetime_charging_time.ipynb: Contains a mini analysis on cycle life vs. charging time.
In addition, we used three scripts for neural net model training in the nn_models directory.
Calls to these scripts look like this:
python mlp_base.py --n_starts 10 --lr 0.001 --rw 0.0001 --hd 10.
- mlp_base.py: Trains MLP models
- cnn_base.py: Trains CNN models without baseline subtraction
- cnn_base_nosubtract.py: Trains CNN models with baseline subtraction
These scripts are used by another script that is written in a proprietary language used to interact with a cluster at IBM; this additional script is not included.
Finally, we include three notebooks that analyze the results of MLP & CNN training
in the nn_models directory.
These notebooks will not run completely without all .pkl or .pt files present.
- Analyze_Results_MLP_CV.ipynb: Analyzes the results from the MLP cross-validation study and selects the best hyperparameters. Also produces intermediate outputs including mlp_predictions.json, mlp_predictions_cycavg.json, and mlp_shap_results.csv.
- MLP_FinalModel.ipynb: Runs the final model based on the results of CV and generates the MLP interpretability figure using shap.
-Analyze_Results_CNN.ipynb: Analyzes the results from the CNN study
(not a formal CV study) and selects the best hyperparameters. The best performing model (CNN_n20000_rw0.001_lr0.0001_do0.0.pt)
is included for reference.
Owner
- Name: Peter Attia
- Login: petermattia
- Kind: user
- Location: Somerville, MA
- Company: Glimpse
- Website: https://petermattia.com
- Repositories: 18
- Profile: https://github.com/petermattia
I ❤️🔋
GitHub Events
Total
- Watch event: 4
- Fork event: 1
Last Year
- Watch event: 4
- Fork event: 1
Committers
Last synced: about 2 years ago
Top Committers
| Name | Commits | |
|---|---|---|
| petermattia | p****a@g****m | 32 |
| dependabot[bot] | 4****] | 7 |
Issues and Pull Requests
Last synced: about 2 years ago
All Time
- Total issues: 5
- Total pull requests: 8
- Average time to close issues: 23 days
- Average time to close pull requests: 4 days
- Total issue authors: 4
- Total pull request authors: 1
- Average comments per issue: 1.4
- Average comments per pull request: 0.25
- Merged pull requests: 7
- Bot issues: 0
- Bot pull requests: 8
Past Year
- Issues: 2
- Pull requests: 1
- Average time to close issues: about 1 month
- Average time to close pull requests: 4 days
- Issue authors: 1
- Pull request authors: 1
- Average comments per issue: 0.5
- Average comments per pull request: 2.0
- Merged pull requests: 0
- Bot issues: 0
- Bot pull requests: 1
Top Authors
Issue Authors
- myalos (2)
- sautee (1)
- fingertap (1)
- Heyuong (1)
Pull Request Authors
- dependabot[bot] (8)