| _version_ | 1866901275106344960 |
|---|---|
| author | Yassine, Benachour |
| author_facet | Yassine, Benachour |
| contents | <h3>What's in this release</h3> <ul> <li>Reproducible preprocessing pipeline (z-score, moving-average smoothing)</li> <li>Sliding-window segmentation: <strong>2.56 s</strong> & <strong>1.0 s</strong> (50% overlap)</li> <li>Feature extraction: time/frequency descriptors + <strong>jerk</strong> and <strong>magnitude</strong></li> <li>Feature sets: <strong>Set_1 (432)</strong>, <strong>Set_2 (576)</strong>, <strong>Set_3 (720)</strong>; final <strong>46-feature</strong> subset (selection scripts/notebooks)</li> <li>Notebooks: <code>0_YB_EDA.ipynb</code>, <code>2_Feature_Extraction.ipynb</code></li> <li>Config-driven paths & parameters (<code>configs/default.yaml</code>)</li> <li>Basic CI and env files (<code>requirements.txt</code>, <code>environment.yml</code>)</li> </ul> <h3>Data location</h3> <p>Dataset is hosted on <strong>Mendeley Data</strong>: DOI <code>10.17632/XXXXXX.1</code><br> (Replace with your issued DOI; large data are not stored in this repo.)</p> <h3>Quick start</h3> <pre><code># conda conda env create -f env/environment.yml conda activate upperlimb # or pip python -m venv .venv && source .venv/bin/activate # Windows: .venv\Scripts\activate pip install -r env/requirements.txt # run bash scripts/01_preprocess.sh bash scripts/02_extract_features.sh bash scripts/03_select_features.sh </code></pre> |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_17164809 |
| institution | Zenodo |
| language | eng |
| publishDate | 2025 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | ybenachour/smartwatch-imu-upperlimb: HemiPhysioData Feature Engineering Pipeline v1.0.0 Yassine, Benachour smartwatch IMU upper-limb rehabilitation human activity recognition feature engineering sliding-window PCA t-SNE Apple Watch inertial sensors Python <h3>What's in this release</h3> <ul> <li>Reproducible preprocessing pipeline (z-score, moving-average smoothing)</li> <li>Sliding-window segmentation: <strong>2.56 s</strong> & <strong>1.0 s</strong> (50% overlap)</li> <li>Feature extraction: time/frequency descriptors + <strong>jerk</strong> and <strong>magnitude</strong></li> <li>Feature sets: <strong>Set_1 (432)</strong>, <strong>Set_2 (576)</strong>, <strong>Set_3 (720)</strong>; final <strong>46-feature</strong> subset (selection scripts/notebooks)</li> <li>Notebooks: <code>0_YB_EDA.ipynb</code>, <code>2_Feature_Extraction.ipynb</code></li> <li>Config-driven paths & parameters (<code>configs/default.yaml</code>)</li> <li>Basic CI and env files (<code>requirements.txt</code>, <code>environment.yml</code>)</li> </ul> <h3>Data location</h3> <p>Dataset is hosted on <strong>Mendeley Data</strong>: DOI <code>10.17632/XXXXXX.1</code><br> (Replace with your issued DOI; large data are not stored in this repo.)</p> <h3>Quick start</h3> <pre><code># conda conda env create -f env/environment.yml conda activate upperlimb # or pip python -m venv .venv && source .venv/bin/activate # Windows: .venv\Scripts\activate pip install -r env/requirements.txt # run bash scripts/01_preprocess.sh bash scripts/02_extract_features.sh bash scripts/03_select_features.sh </code></pre> |
| title | ybenachour/smartwatch-imu-upperlimb: HemiPhysioData Feature Engineering Pipeline v1.0.0 |
| topic | smartwatch IMU upper-limb rehabilitation human activity recognition feature engineering sliding-window PCA t-SNE Apple Watch inertial sensors Python |
| url | https://doi.org/10.5281/zenodo.17164809 |