AI-based Sentiment Analysis for Hotel Customer Reviews

An end-to-end solution for analyzing hotel customer reviews using artificial neural networks (ANN) and Ordinary Least Squares (OLS) regression. This platform provides a comprehensive ecosystem that enables researchers to advance sentiment analysis capabilities and allows developers to build and deploy sentiment analysis applications.

🎯 Project Overview

The platform focuses on efficient sentiment analysis through:

• Situational model application
• Model training and validation
• Model refinement

Implementation is handled via Over-The-Air deployment of:

1. AI/OLS models (knowledge base)
2. Model activations (activation base)
3. Training material (learning base)
4. Analysis routines (codeBase)

📊 Data Source

Our analysis is based on Booking.com reviews for hotels in Spain, comprising approximately 950 reviews. The collected features include:

• Customer review scores (1-10 scale)
• Review text classifications
• Free cancellation information
• Review counts and price metrics

🚀 Model Performance

Neural Network (TensorFlow)

• R² score: 0.875 (87.5% variance explained)
• RMSE: 0.212 (average deviation of 0.21 points on 10-point scale)
• MAE: 0.161 (predictions off by 0.16 points on average)

OLS Model

• R² score: 0.576 (57.6% variance explained)
• RMSE: 0.391
• MAE: 0.170

Comparative Analysis

The Neural Network outperforms the OLS model due to its ability to capture non-linear relationships. While both models maintain consistent prediction accuracy (similar MAE), the Neural Network's lower RMSE indicates fewer extreme prediction errors.

🛠️ Installation

Prerequisites

pip install tensorflow==2.15.0
pip install pandas==2.1.4
pip install numpy==1.26.3
pip install matplotlib==3.8.2
pip install seaborn==0.13.1
pip install scikit-learn==1.3.2
pip install statsmodels==0.14.1

Docker Setup

1. Install Docker Desktop from the official website
2. Create required volume:

docker volume create ai_system

3. Pull Docker Images:

docker pull wasif89/knowledgebase_sentiment_analysis:latest
docker pull wasif89/codespace_sentiment_analysis:latest

📁 Project Structure

Code Organization

code/
├── scrape_reviews.py    # Data collection script
├── data_clean.ipynb    # Preprocessing notebook
├── train_model.py      # Neural network implementation
├── train_ols.py        # OLS model implementation

Data Organization

data/
├── hotels_list.csv            # Raw data
├── joint_data_collection.csv  # Processed data

Docker Images

images/
├── knowledgeBase_sentiment_analysis/  # Model container
├── codeBase_sentiment_analysis/       # Data container

🔍 Model Details

Features Used

• review_text: Text sentiment (0-4)
• free_cancellation: Binary flag
• reviews_count_scaled: Normalized count
• price_scaled: Normalized price

Running Analysis

For AI analysis:

cd scenarios/apply_ai_sentiment_analysis
docker-compose up

For OLS analysis:

cd ../apply_ols_sentiment_analysis
docker-compose up

🐳 Docker Implementation

Docker Compose Configuration

version: '3'
services:
  knowledgebase:
    image: wasif89/knowledgebase_sentiment_analysis:latest
    volumes:
      - ai_system:/tmp

  codespace:
    image: wasif89/codespace_sentiment_analysis:latest
    volumes:
      - ai_system:/tmp

volumes:
  ai_system:
    external: true

📝 Course Information

This repository is created and maintained by Mustafa Wasif as part of the course 'M. Grum: Advanced AI-based Application Systems' by the Junior Chair for Business Information Science, esp. AI-based Application Systems at University of Potsdam.

📄 License

This project is licensed under the AGPL-3.0 license.