Apple Quality Classification

A machine learning project that uses Logistic Regression to classify apples as good or bad based on various features. The dataset is preprocessed, trained with models using L2 regularization, and evaluated using performance metrics and an ROC curve.

Features

• 📌 Data Preprocessing: Cleans missing values and applies feature scaling.
• 📌 Logistic Regression: Implements classification with and without regularization.
• 📌 Model Comparison: Evaluates multiple models with different regularization strengths.
• 📌 Performance Metrics: Calculates accuracy, precision, recall, F1-score, and ROC-AUC.
• 📌 ROC Curve Visualization: Plots the best model’s performance.

---

🖼️ Screenshots

🔹 ROC Curve

---

📂 Dataset

The dataset is publicly available and can be accessed here:
📌 Apple Quality Dataset

Features:

• Various numerical attributes related to apple quality.
• Target variable: Quality (good = 1, bad = 0).

---

🛠️ Technologies & Libraries

Library	Purpose
pandas	Data processing & cleaning
sklearn.model_selection	Train-test splitting
sklearn.preprocessing	Feature scaling
sklearn.linear_model	Logistic Regression models
sklearn.metrics	Performance evaluation
matplotlib.pyplot	Visualization (ROC Curve)

🏗️ Built With

• 
• 
• 

---

🎯 How to Run

1️⃣ Clone the repository:

git clone https://github.com/your-username/apple-quality-classification.git
cd apple-quality-classification

2️⃣ Install dependencies:

pip install pandas scikit-learn matplotlib

3️⃣ Run the Python script:

python main.py

---

📊 Model Comparison

Model Type	Accuracy	Precision	Recall	F1-score	ROC-AUC
No Regularization	0.85	0.87	0.83	0.85	0.86
L2 (C=0.1)	0.88	0.89	0.86	0.87	0.89
L2 (C=1)	0.90	0.91	0.88	0.89	0.91
L2 (C=10)	0.89	0.90	0.87	0.88	0.90

✔ Best Model: L2 Regularization with C=1, achieving the highest accuracy and balanced precision/recall.