Sentiment Analysis on Hotel Reviews

Project Overview

This project focuses on performing sentiment analysis on hotel reviews. The goal is to collect, clean, and analyze review data to develop a machine learning model that can predict the sentiment (positive, negative, or neutral) of the reviews. The project involves four key stages: data collection, data cleaning, building a sentiment analysis machine learning model and visualizing the dataset in a map

Table of Contents

1. Project Overview
2. Requirements
3. Data Collection
4. Data Cleaning
5. Exploratory Data Analysis (EDA)
6. Model Training
7. Evaluation
8. How to Run
9. File Structure
10. Contributors

Requirements

To run this project, you will need the following libraries installed:

• Python 3.x
• Pandas
• NumPy
• Pickle (Model loading)
• Scikit-learn
• NLTK (Natural Language Toolkit)
• BeautifulSoup (Web scraping)
• Wordcloud
• Keras
• SQLalchemy
• Pymongo
• Matplotlib (for data visualization)

You can install all dependencies by running:

bash
Copy code
pip install -r requirements.txt

Data Collection

Source:

Hotel reviews can be collected from various sources, such as:

1. Web scraping hotel booking platforms (e.g., TripAdvisor, Yelp.com). See 00 Data Collection for details
2. Public datasets from sources like Kaggle (Booking hotel reviews dataset).

Data Cleaning

Raw data may contain noise, missing values, or irrelevant information. The data_cleaning.py script handles data preprocessing, which includes:

• Removing duplicates
• Filling or removing missing values
• Normalizing text (removing punctuation, converting to lowercase)
• Tokenization, stopword removal, and lemmatization using NLTK or SpaCy
• Encoding sentiment labels (e.g., Positive = 1, Negative = 0)

Model Training

For sentiment analysis, several machine learning models can be used, including:

• Logistic Regression
• Support Vector Machine (SVM)
• K-nearest neighbor

The model_training.py script contains the process of training the model. It includes:

1. Vectorization: Use TF-IDF or Word2Vec for text vectorization.
2. Model Training: Train the machine learning model using labeled data.
3. Hyperparameter Tuning: Apply GridSearchCV or RandomSearch for optimizing hyperparameters. Example:

Evaluation

After training, the model is evaluated using various metrics:

• Accuracy
• Precision
• Recall
• F1-Score
• Confusion Matrix

Contributors Hugo Villanueva ([email protected]) Feel free to open an issue or contribute to the project!