# Catalog — Data

来源：https://www.kaggle.com/competitions/catalog/data

本地 Meta Kaggle 快照；不是本次实时页面。

# Dataset Description

There is a directory of the various positions used by TMK. Each position in the directory
contains only two attributes: "Title" and "Group". Objective: to predict the "Group" attribute by
attribute "Title".

The quality metric is accuracy - the percentage of correct predictions.

## Files

*   **train.txt** - a file with a training sample: each line represents one position and consists of the name of the position and the group, separated by a tab character.
*   **test.txt** - a file with a test sample: the file consists of 2346 lines, each line completely consists of the name of the position for which you want to define a group.

## Columns

*   `Название` - product name
*   `Группа` - group