BUPTLdy/Sentiment-Analysis

Chinese Shopping Reviews sentiment analysis

What it solves

This project provides a system for classifying the sentiment of Chinese product reviews as either positive or negative. It serves as both a historical record of early sentiment analysis techniques and a modern, reproducible baseline for sentiment classification.

How it works

The project offers two main implementation paths:

  1. Modern Baseline: A Python 3 compatible CLI tool that uses character-level TF-IDF (Term Frequency-Inverse Document Frequency) to represent text without needing a Chinese word segmentation dictionary, paired with a Linear SVM (Support Vector Machine) for classification.
  2. Legacy Implementations: Historical code using Word2Vec embeddings combined with either RBF SVM or Keras LSTM networks (designed for Python 2.7 and TensorFlow 1).

Who it’s for

Developers and students looking for a simple, reproducible baseline for Chinese sentiment analysis or those researching the evolution of sentiment classification methods from Word2Vec/LSTM to TF-IDF/SVM.

Highlights

  • Dictionary-free: The modern implementation uses character-level TF-IDF, removing the dependency on external Chinese segmentation tools.
  • Reproducible: Includes a CLI for training and predicting with fixed random seeds for consistent evaluation.
  • Historical Context: Preserves legacy implementations for research and comparison purposes.
  • Metadata Tracking: Saves training time, data volume, and accuracy metrics alongside the model.

Related

  • Dispatch
    Detecting LLM-Generated Text with Classical Machine Learning – Results and Insights

    A scikit-learn TF‑IDF + LinearSVC pipeline can distinguish modern LLM‑generated Chinese text from human‑written text with ~85% sentence‑level accuracy, demonstrating that “classical” ML still works for AI‑text detection.

  • Project
    DLLXW/baby-llama2-chinese

    A small-parameter Chinese Llama2 implementation providing a full pipeline for pre-training, SFT, and inference, designed to help beginners learn LLM development.

  • Project
    larashero3-dotcom/lieflat-less-ai-tone

    A linguistic research project and toolset that quantifies "AI flavor" in Chinese text and provides a set of rules to remove these markers for more human-like writing.

  • Dispatch
    Getting Started with Sentiment Analysis on Twitter

    Hugging Face provides a guide on using the Inference API to perform automated sentiment analysis on Twitter data for both developers and non-coders.

  • Project
    QwenAudio/Fun-ASR

    A family of high-precision end-to-end speech recognition models optimized for noisy environments, diverse Chinese dialects, and high-throughput production deployment.