##  Dataset description

The participants of this competition are provided with the BookSCE dataset. 
 The dataset consists of around 11K books accompanied by the literary epochs label:  (1) Romanticism (1798-1837), (2) Victorian Literature (1837-1901), (3) Modernism (1900-1945), (4) Postmodernism (1945-2000), and (5) our days (from 2000).

## Ground truth
To train the models, the participants are provided with a train set that consists of books in the "txt" format accompanied by the corresponding ground truth file "train.scv". The ground truth includes the book IDs and the corresponding epoch labels:
- Romanticism
- Victorian
- Modernism 
- PostModernism
- OurDays

The classifier's input is a book in text format, and its output is a single value (epoch). 
The training and validation sets are released at the beginning of the competition. The test set will be released (without labels) a week before the deadline. 
The results will be evaluated and ranked based on average accuracy.

## Files

*   **train.zip** - the training set files. 
*   **train.csv** - the training set labels
*   **valid.zip** - the validation set files
*   **valid.csv** - the validation set labels
*   **test.zip** - the test set files (to be released on July 10, 2023)
*   **sample_submission.csv** - a sample submission file in the correct format

**All the files are given in the form of 1000-word chunks.**

## Columns

*   `BOOK_id` - id of a book
*   `Epoch` - is a label (one of 5 classes described above)