Sensor data from smartphones worn by 30 participants performing activities like walking, sitting, and standing. The dataset was created by researchers for machine learning benchmarks and is hosted by the UCI Machine Learning Repository. The original collection date is unspecified.
Use Cases
- Classify activities like walking, sitting, and standing using triaxial accelerometer and gyroscope time-series data.
- Detect postural transitions such as sit-to-stand using sensor signal patterns and derived features.
- Train models to segment continuous sensor streams into discrete activity labels.
- Benchmark feature engineering and deep learning approaches for time-series classification on inertial data.
Strengths
- Data collected from 30 participants, providing a multi-subject basis for model generalization.
- Includes both basic activities and postural transitions, covering a wider range of human movements.
Limitations
- The exact number of records and sensor sampling rate are unspecified, complicating reproducibility.
- Lack of detailed participant demographics (age, gender) limits analysis of population-specific models.
- Unknown collection date and environment may reduce relevance for modern smartphone sensor configurations.
Provenance
- Source
- UCI Machine Learning Repository
- Collection Method
- Data gathered from smartphone sensors (accelerometer, gyroscope) worn by participants during scripted activities.
- Time Range
- null
- Freshness
- null
- Geography
- null