Milk-Quality-Classification-ML

Project Overview

This project applies supervised machine learning to classify milk quality based on seven chemical and sensory features. We implement and compare Logistic Regression, Decision Trees, and Random Forests, evaluating their performance through cross-validation, hyperparameter tuning, and normalization. Additionally, we explore TabR (2023), a hybrid deep learning model that integrates k-Nearest Neighbors (KNN) for tabular data classification.

Dataset

• Source: Kaggle
• Size: 1,059 entries
• Missing data: 0
• Duplicated data: 976
• Features:
  • pH
  • Temperature
  • Taste
  • Odor
  • Fat
  • Turbidity
  • Color
• Target Variable: Milk Quality (Low, Medium, High)

Methodology

1. Data Preprocessing

   • Data cleaning and handling duplicates
   • Exploratory Data Analysis (EDA)
   • Train-test split (80-20)
   • Normalization

2. Model Selection & Training

   • Logistic Regression (Regularization tuning)
   • Decision Tree (Depth optimization)
   • Random Forest (Number of trees tuning)
   • TabR (Hybrid KNN-Deep Learning model)

3. Hyperparameter Tuning

   • Manual tuning with 5-fold cross-validation
   • Evaluated C (Logistic Regression), max_depth (Decision Tree), and n_estimators (Random Forest)

4. Evaluation Metrics

   • Accuracy
   • Feature importance analysis
   • Impact of normalization

Key Findings

• Random Forest achieved the highest accuracy (99.88%), outperforming other models due to ensemble learning.
• Normalization significantly improved Logistic Regression accuracy by 17.4%.
• Tree-based models (Decision Tree, Random Forest) performed best, prioritizing pH as the most important predictor.
• TabR provided insights into retrieval-augmented deep learning for tabular data.

Technologies Used

• Python (Pandas, NumPy, Scikit-learn, Seaborn, Matplotlib)
• Machine Learning Models (Logistic Regression, Decision Tree, Random Forest)
• Deep Learning Approach (TabR, KNN-based feature augmentation)

References

1. Hastie, T., Tibshirani, R., & Friedman, J. (2009). The Elements of Statistical Learning: Data Mining, Inference, and Prediction (2nd ed.). Springer.
2. Gorishniy, Y., Rubachev, I., Kartashev, N., Shlenskii, D., Kotelnikov, A., & Babenko, A. (2023). TabR: Tabular Deep Learning Meets Nearest Neighbors in 2023. arXiv preprint arXiv:2307.14338.

Repository Link

GitHub Repository