Fake-review-detector-distilbert-V2

Fake Review Detection (DistilBERT)

Overview

This model performs fake review detection on e-commerce customer reviews.

It predicts whether a review is authentic (real) or fake (spam/manipulated).

The model is based on the DistilBERT transformer architecture and fine-tuned on an Amazon-style review dataset.

This model is designed as part of a business-oriented application to improve trust and transparency in online marketplaces.

---

Model Details

• Base model: DistilBERT (distilbert-base-uncased)
• Task: Text Classification (Fake Review Detection)
• Framework: Hugging Face Transformers

---

Labels

• LABEL_0 → Real Review ✅
• LABEL_1 → Computer-Generated Review ❌

---

Dataset

Features:

• text: customer review content
• label:
  • 0 → Real(verified) Review
  • 1 → Computer-Generated(Fake) review

A stratified subset of 10,000 samples was used to improve training efficiency while maintaining label balance.

Data split:

• Train: 8,000
• Validation: 1,000
• Test: 1,000

---

Evaluation Results

The model achieves a good balance between performance and efficiency, making it suitable for real-time applications.

---

Usage

from transformers import pipeline

classifier = pipeline(
    "text-classification",
    model="JerryJJJJJ/amazon-fake-review-detector-distilbert"
)

classifier("Buy now!!! Best product ever!!! Limited offer!!!")