Stroke case - Smart Belt Dataset

The dataset contains 22,056,714 samples collected from patients with stroke. The data was gathered using a smart belt with accelerometer and gyroscope sensors. Two belts were used to record the physical activities of the patients, where Belt 1 (B1) comprised 3,158,873 samples, and Belt 2 (B2) contained 18,897,841 samples. Each belt was attached at specific body locations around the waist, including the left, right, and middle hip. The variables generated from the sensors were recorded at a frequency of 100Hz and are described as follows:

1. Packet ID: distinguishes data packets that change with time.

2. Timestamp: the time that a reading was taken on a sensor

3. Timer: Timer of the sensor

4. Acceleration X: Linear acceleration in the X-axis

5. Acceleration Y: Linear accelerations in the y-axis

6. Acceleration Z: Linear acceleration in the z-axis

7. Gyroscope X: Angular acceleration in the x-axis

8. Gyroscope Y: Angular acceleration in the y-axis

9. Gyroscope Z: Angular acceleration in the z-axis

10. Temperature: internal temperature for signal compensation

After acquiring raw sensor signals, a series of minor preprocessing tasks were performed on this dataset. These include reformatting the data with a suitable format, removing null values, and preparing the data for analysis.