# DL for Tabular Playground Series - Apr 2022 — Data

来源：https://www.kaggle.com/competitions/dl-for-tabular-playground-series-apr-2022/data

采集时间：2026-10-19T08:18:39Z

In this competition, you'll classify 60-second sequences of sensor data, indicating whether a subject was in either of two activity states for the duration of the sequence.

## Files and Field Descriptions ##

- **train.csv** - the training set, comprising ~26,000 60-second recordings of thirteen biological sensors for almost one thousand experimental participants
  - `sequence` - a unique id for each sequence
  - `subject` - a unique id for the subject in the experiment
  - `step` - time step of the recording, in one second intervals
  - `sensor_00` - `sensor_12` - the value for each of the thirteen sensors at that time step
- **train_labels.csv** - the class label for each sequence. 
  - `sequence` - the unique id for each sequence.
  - `state` - the state associated to each sequence. This is the target which you are trying to predict.
- **test.csv** - the test set. For each of the ~12,000 sequences, you should predict a value for that sequence's `state`.
- **sample_submission.csv** - a sample submission file in the correct format.