Amazon Reviews Sentiment Analysis Model

Model Description

This model is a sentiment analysis model fine-tuned using BertForSequenceClassification on the Amazon Reviews dataset. It classifies Amazon product reviews into sentiment categories: negative, neutral, or positive. Intended for research, educational, and non-commercial use only.

Base Model

bert-base-uncased
Architecture: Transformer (BERT)
Head: Sequence Classification

Intended Use

✅ Allowed Uses

Academic research
Educational projects
Personal learning
Non-commercial applications
Experiments and benchmarking

❌ Prohibited Uses

Commercial use
Selling or reselling the model
Paid APIs or SaaS products
Monetized applications or services

Commercial use is strictly prohibited under the CC BY-NC 4.0 license.

Training Data

Trained on the Amazon Reviews dataset:

Language: English
Domain: E-commerce product reviews
Data type: Text reviews with sentiment labels

The original dataset creators retain all rights to the data. Users should consult the dataset’s original license for details.

Training Procedure

Model: BertForSequenceClassification
Framework: Hugging Face Transformers
Number of labels: 3
Loss Function: Cross-entropy loss
Training performed on GPU if available, otherwise CPU

Label Mapping

Label ID	Sentiment
0	Negative
1	Neutral
2	Positive

Evaluation

Evaluated using a multi-class classification report with three categories:

Negative
Neutral
Positive

Metrics include precision, recall, F1-score, and support (per class). Performance may vary depending on product category and review style.

Limitations and Bias

Reflects biases in Amazon reviews
May not perform well on non-product text
Not suitable for non-English languages
Predictions are subjective, not factual judgments

Ethical Considerations

Analyze subjective content only; not for high-stakes decisions.

How to Use

from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

model_name = "mianzaka/sentiment-analysis-model"

tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSequenceClassification.from_pretrained(model_name)

text = "The product quality is decent but delivery was slow."
inputs = tokenizer(text, return_tensors="pt", truncation=True, padding=True)

with torch.no_grad():
    outputs = model(**inputs)

predicted_label = torch.argmax(outputs.logits, dim=1).item()
label_map = {0: "Negative", 1: "Neutral", 2: "Positive"}
print("Predicted sentiment:", label_map[predicted_label])

License

Released under CC BY-NC 4.0. Commercial use, resale, or monetization is prohibited. Full license: https://creativecommons.org/licenses/by-nc/4.0/

Citation

@misc{sentiment-analysis-model,
  author = {Mian Zaka},
  title  = {Amazon Reviews Sentiment Analysis Model},
  year   = {2026},
  publisher = {Hugging Face}
}

Contact

For questions or feedback, contact the model author via Hugging Face.

Downloads last month: 15

Safetensors

Model size

0.1B params

Tensor type

F32