# Data Description

This page provides information about the dataset for the Bangla Sentiment Analysis Competition.

## Files
* **train.csv** - the training dataset containing Bangla text examples with sentiment labels
* **sample_submission.csv** - a sample submission file in the correct format

## Dataset Overview

The dataset consists of 103 Bangla language text samples related to banking and company services, with sentiment annotations.


## Notes on the Data

1. **Text Content**: The text content includes customer comments, reviews, queries, and feedback about banking services, mobile apps, customer support experiences, product quality, and related topics.

2. **Language**: All text is in Bangla (Bengali) language. Proper UTF-8 encoding should be used to correctly process the text.

3. **Pre-processing**: No pre-processing has been applied to the text - you may need to handle punctuation, special characters, and normalization.

4. **Task**: The core task is multi-class classification to determine whether a given Bangla text expresses a positive, negative, or neutral sentiment.

## Prediction Task

For each text in the test set, you must predict whether it expresses a positive, negative, or neutral sentiment.
Learn more about this competition: https://youtu.be/4n1algYgx5Y