Datasets:
Formats:
parquet
Size:
100M - 1B
ArXiv:
Tags:
timeseries
timeseries clustering
changepoint-detection
correlation-structure
Synthetic
benchmark
License:
Added usage examples
Browse files
README.md
CHANGED
@@ -467,7 +467,60 @@ Our accompanying paper provides complete methodological details, baseline findin
|
|
467 |
GitHub codebase includes the generation, validation and use case code and is configured to automatically load the data.
|
468 |
|
469 |
## Usage Guidance
|
470 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
471 |
|
472 |
## Authors
|
473 |
- Isabella Degen, University of Bristol
|
|
|
467 |
GitHub codebase includes the generation, validation and use case code and is configured to automatically load the data.
|
468 |
|
469 |
## Usage Guidance
|
470 |
+
### Configuration Concept
|
471 |
+
The configuration follows the convention: `<generation_stage>_<completeness_level>_<file_type>` and allows
|
472 |
+
access to a specific subset of the data.
|
473 |
+
|
474 |
+
Possible values are:
|
475 |
+
|
476 |
+
Generation Stages:
|
477 |
+
- **raw**: raw data, segmented but not correlated
|
478 |
+
- **correlated**: correlated data according to a specific correlation strtucture, normal distributed
|
479 |
+
- **nonnormal**: distribution shifted, correlated data
|
480 |
+
- **downsampled**: resampled non-normal data from 1s to 1min
|
481 |
+
|
482 |
+
Completeness Levels:
|
483 |
+
- **complete**: 100% of the data
|
484 |
+
- **partial**: 70% of the data (30% of observations dropped at random)
|
485 |
+
- **sparse**: 10% of the data (90% of observations dropped at random)
|
486 |
+
|
487 |
+
**File Type**
|
488 |
+
- **data**: loads the times series data file (needed for training algorithms)
|
489 |
+
- **labels**: loads the labels file for the ground truth (perfect) segmentation and clustering (needed for validating the results)
|
490 |
+
- **badclustering_labels**: loads the labels file for a degraded clustering with controlled segmentation and/or cluster assignment mistakes
|
491 |
+
|
492 |
+
### Splits
|
493 |
+
The main splits are:
|
494 |
+
- **exploratory**: for experimentation and training
|
495 |
+
- **confirmatory**: for testing and validation
|
496 |
+
Consider that depending on the application and study design, a single subject might be sufficient for training.
|
497 |
+
|
498 |
+
Additional splits are:
|
499 |
+
- **reduced_11_clusters**(_exploratory or _confirmatory): same data including 11 of the original 23 clusters (selected at random)
|
500 |
+
- **reduced_6_clusters**(_exploratory or _confirmatory): same data including t of the original 23 clusters (selected at random)
|
501 |
+
- **reduced_50_segments**(_exploratory or _confirmatory): same data including 50 of the original 100 segments (selected at random)
|
502 |
+
- **reduced_25_segments**(_exploratory or _confirmatory): same data including 25 of the original 100 segments (selected at random)
|
503 |
+
|
504 |
+
### Quick Start
|
505 |
+
#### Example 1 - complete and correlated data variant
|
506 |
+
1. Load the data for all 30 exploratory subjects for the complete and correlated data variant into pandas df:
|
507 |
+
```python
|
508 |
+
import pandas as pd
|
509 |
+
from datasets import load_dataset
|
510 |
+
correlated_data = load_dataset("idegen/csts", name="correlated_complete_data", split="exploratory")
|
511 |
+
df_correlated = correlated_data.to_pandas()
|
512 |
+
df_correlated.head()
|
513 |
+
```
|
514 |
+
2. Load the ground truth labels for these subjects
|
515 |
+
```python
|
516 |
+
import pandas as pd
|
517 |
+
from datasets import load_dataset
|
518 |
+
correlated_labels = load_dataset("idegen/csts", name="correlated_complete_labels", split="exploratory")
|
519 |
+
df_correlated_labels = correlated_labels.to_pandas()
|
520 |
+
df_correlated_labels.head()
|
521 |
+
```
|
522 |
+
|
523 |
+
... more examples coming soon
|
524 |
|
525 |
## Authors
|
526 |
- Isabella Degen, University of Bristol
|