🍽️ Recipe Reviews Sentiment Analysis (Data Science with Machine Learning Project)

📌 Project Overview

This project performs sentiment analysis on recipe reviews using machine learning and natural language processing (NLP). User reviews are classified into positive, neutral, or negative sentiment based on review text and star ratings.The dataset I got from here: https://archive.ics.uci.edu/ . It is a classification dataset containing user reviews, ratings, and recipe information.

The project demonstrates an end‑to‑end machine learning pipeline, including data cleaning, feature engineering, AutoML, and hyperparameter optimization.

---

📊 Dataset Description

The dataset contains:

• User review text
• Star ratings (1–5)
• Recipe names
• User metadata

Data Cleaning Steps

• Removed unnecessary identifier columns
• Dropped rows with missing review text or ratings
• Converted star ratings into sentiment labels:
  • 1–2 stars → Negative
  • 3 stars → Neutral
  • 4–5 stars → Positive

---

🎯 Machine Learning Objective

Multiclass classification to predict review sentiment:

• Positive
• Neutral
• Negative

---

🔍 Feature Engineering

The model uses:

• Numerical features (scaled with StandardScaler)
• Categorical feature (recipe_name) using OneHotEncoder
• Text feature (review text) using TF‑IDF Vectorization

---

✂️ Train–Test Split

• 80% Training data
• 20% Testing data
• Stratified sampling to preserve class balance

---

🤖 AutoML (AutoGluon)

AutoGluon was used to:

• Automatically train and compare multiple models
• Select the best performing model
• Optimize accuracy without manual tuning

Evaluation Metric: Accuracy
Problem Type: Multiclass Classification

---

⚙️ Logistic Regression + Optuna Optimization

A custom pipeline was built using:

• ColumnTransformer
• Logistic Regression
• TF‑IDF Vectorizer

Optimized Hyperparameters

• Regularization strength (C)
• Maximum TF‑IDF features

Optuna was used with cross‑validation to find the best parameters.

---

📈 Model Evaluation

Evaluation was performed using:

• Accuracy score
• Precision, Recall, F1‑Score
• Confusion Matrix visualization

---

🧰 Tools & Libraries

• Python
• Pandas & NumPy
• Scikit‑learn
• AutoGluon
• Optuna
• Matplotlib

---

💡 Skills Demonstrated

• Natural Language Processing (NLP)
• Machine Learning Pipelines
• AutoML
• Hyperparameter tuning
• Data preprocessing
• Model evaluation
• GitHub documentation

---

🚀 How to Run the Project

1. Clone the repository
2. Install dependencies using requirements.txt
3. Run the notebook or Python script

---

✅ Status

✔️ Completed