The project aims to utilize mobile device accelerometer data to detect and categorize user-defined actions in real-time, ensuring accurate and timely identification.
-
Updated
May 22, 2026 - Python
#
wisdm
Here are 5 public repositories matching this topic...
Subject-independent HAR on WISDM (UCI 507). Session-safe 20 Hz repair, GroupKFold XGBoost, and a CPU FastAPI. Leaky phone flatten 0.89 macro-F1 drops to 0.29 with subject holdout; watch statistical XGBoost is 0.70.
docker machine-learning time-series xgboost human-activity-recognition onnx mlflow fastapi wisdm wearable-sensors
-
Updated
Aug 22, 2026 - Jupyter Notebook
-
Updated
Feb 4, 2019
Human Activity Recognition on the WISDM dataset using a vanilla Transformer (PyTorch) with an interactive Plotly Dash dashboard. 73.77% val accuracy with 4-sensor fusion.
-
Updated
May 27, 2026 - Python
Real-time human activity recognition: SVM vs LSTM on tri-axial accelerometer data, with Flask inference API, React dashboard, and ESP32 firmware.
react flask machine-learning firebase typescript deep-learning tensorflow svm esp32 scikit-learn accelerometer lstm human-activity-recognition wisdm
-
Updated
May 4, 2026 - Jupyter Notebook
Add this topic to your repo
To associate your repository with the wisdm topic, visit your repo's landing page and select "manage topics."