metadata

datasets:
  - sentiment-analysis-dataset
language:
  - en
task_categories:
  - text-classification
task_ids:
  - sentiment-classification
tags:
  - sentiment-analysis
  - text-classification
  - balanced-dataset
  - oversampling
  - csv
pretty_name: Sentiment Analysis Dataset (Imbalanced)
dataset_info:
  features:
    - name: text
      dtype: string
    - name: label
      dtype: int64
  splits:
    - name: train
      num_examples: 83989
    - name: validation
      num_examples: 10499
    - name: test
      num_examples: 10499
  format: csv

Sentiment Analysis Dataset

Overview

This dataset is designed for sentiment analysis tasks, providing labeled examples across three sentiment categories:

0: Negative
1: Neutral
2: Positive

It is suitable for training, validating, and testing text classification models in tasks such as social media sentiment analysis, customer feedback evaluation, and opinion mining.

Dataset Details

Key Features

Type: CSV
Language: English
Labels:
- 0: Negative
- 1: Neutral
- 2: Positive
Pre-processing:
- Duplicates removed
- Null values removed
- Cleaned for consistency

Dataset Split

Split	Rows
Train	83,989
Validation	10,499
Test	10,499

Format

Each row in the dataset consists of the following columns:

text: The input text data (e.g., sentences, comments, or tweets).
label: The corresponding sentiment label (0, 1, or 2).

Usage

Installation

Download the dataset from the Hugging Face Hub or your preferred storage location.

Loading the Dataset

Using Pandas

import pandas as pd

# Load the train dataset
train_df = pd.read_csv("path_to_train.csv")
print(train_df.head())

# Columns: text, label

Using Hugging Face's `datasets` Library

from datasets import load_dataset

# Load the dataset
dataset = load_dataset("your-dataset-path")

# Access splits
train_data = dataset["train"]
validation_data = dataset["validation"]
test_data = dataset["test"]

# Example: Printing a sample
print(train_data[0])

Example Usage

Here’s an example of using the dataset to fine-tune a sentiment analysis model with the Hugging Face Transformers library:

from transformers import AutoTokenizer, AutoModelForSequenceClassification, Trainer, TrainingArguments
from datasets import load_dataset

# Load dataset
dataset = load_dataset("your-dataset-path")

# Load model and tokenizer
model_name = "bert-base-uncased"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSequenceClassification.from_pretrained(model_name, num_labels=3)

# Tokenize dataset
def tokenize_function(examples):
    return tokenizer(examples["text"], padding="max_length", truncation=True)

tokenized_datasets = dataset.map(tokenize_function, batched=True)

# Prepare training arguments
training_args = TrainingArguments(
    output_dir="./results",
    evaluation_strategy="epoch",
    save_strategy="epoch",
    learning_rate=2e-5,
    per_device_train_batch_size=16,
    num_train_epochs=3,
    weight_decay=0.01,
    load_best_model_at_end=True,
)

# Initialize Trainer
trainer = Trainer(
    model=model,
    args=training_args,
    train_dataset=tokenized_datasets["train"],
    eval_dataset=tokenized_datasets["validation"],
)

# Train model
trainer.train()

Applications

This dataset can be used for:

Social Media Sentiment Analysis: Understand the sentiment of posts or tweets.
Customer Feedback Analysis: Evaluate reviews or feedback.
Product Sentiment Trends: Monitor public sentiment about products or services.

License

This dataset is released under the [Insert Your Chosen License Here]. Ensure proper attribution if used in academic or commercial projects.

Citation

If you use this dataset, please cite it as follows:

@dataset{your_name_2024,
  title        = {Sentiment Analysis Dataset},
  author       = {Syed Khalid Hussain},
  year         = {2024},
  url          = {https://huggingface.co/datasets/syedkhalid076/Sentiment-Analysis}
}

Acknowledgments

This dataset was curated and processed by Syed Khalid Hussain. The author takes care to ensure high-quality data, enabling better model performance and reproducibility.

Author: Syed Khalid Hussain