English | 简体中文

Sentiment_Analysis_BERT

📌 Project Overview

This project is based on the Chinese pre-trained language models bert-base-chinese and xlm-roberta-base, combined with the RNN neural network structure, to build a lightweight Chinese sentiment analysis system. The model is trained on the ChnSentiCorp dataset and supports rapid deployment and use.

The project is suitable for Chinese short-text sentiment classification tasks and can be easily extended to scenarios like comment analysis and user feedback recognition.

🎯 Experimental Results

The model was trained with 1000 training samples from the ChnSentiCorp dataset for 3 epochs. The accuracy on the full test set is shown below:

Model Architecture	Accuracy (ACC)
BERT (bert-base-chinese)	88.42%
XLM-RoBERTa (xlm-roberta-base)	87.75%
BERT + RNN	88.92%
XLM-RoBERTa + RNN	88.67%

🚀 Quick Start

1. Install dependencies:

pip install -r requirements.txt

1. Run the demo:

python demo.py

⚙️ Model Training and Configuration

Start the training task using main.py. You can modify training parameters (such as model type, number of epochs, batch size, etc.) in config.py:

python main.py --model bert+rnn --num_epoch 5

The trained model will be saved in the ./trained_model directory (the trained model has already been uploaded and can be used directly).

📈 Model Evaluation and Deployment

Test the model performance on the validation set:

python test_set.py

Use the trained model to predict sentiment:

python demo.py

You can input any Chinese sentence, and the model will automatically determine its sentiment (positive/negative).

📂 Directory Structure

├── config.py               # Hyperparameter configuration
├── main.py                 # Model training entry point
├── demo.py                 # Single-sentence sentiment prediction
├── test_set.py             # Evaluate the model on the test set
├── model.py                # Model definition (including RNN structure)
├── trained_model/          # Directory for saving trained models
├── requirements.txt        # Python dependencies