README.md

Machine Learning Model for Classification

This repository contains a machine learning pipeline that processes a dataset, trains multiple models, tunes hyperparameters, and evaluates performance using accuracy and confusion matrices.

The pipeline supports handling imbalanced data using SMOTE and includes Random Forest hyperparameter tuning using GridSearchCV.

🚀 Features

• Preprocessing: Encodes categorical data, normalizes numerical features, and handles missing values.
• Handles Imbalanced Data: Uses SMOTE (Synthetic Minority Over-sampling Technique).
• Trains multiple models:
  • Decision Tree
  • Random Forest
  • Support Vector Machine (SVM)
  • k-Nearest Neighbors (KNN)
  • (Optional) LightGBM (if installed)
• Hyperparameter Tuning: Optimizes Random Forest with GridSearchCV.
• Feature Importance Analysis: Displays feature contributions for model decisions.
• Model Saving and Loading: Saves the best-trained model for future predictions.

📂 Dataset Information

Predicting Diabetes based on factors such as bmi, age, number of pregnancies etc

For Each Attribute:

Number of times pregnant Plasma glucose concentration a 2 hours in an oral glucose tolerance test Diastolic blood pressure (mm Hg) Triceps skin fold thickness (mm) 2-Hour serum insulin (mu U/ml) Body mass index (weight in kg/(height in m)^2) Diabetes pedigree function Age (years)

The dataset should be a CSV file with relevant attributes for classification tasks. The dataset used in this project focuses on medical diagnostics, predicting diabetes based on health indicators.

Dataset Columns

• Pregnancies: Number of times the person has been pregnant
• Glucose: Plasma glucose concentration after a 2-hour oral glucose tolerance test
• BloodPressure: Diastolic blood pressure (mm Hg)
• SkinThickness: Triceps skin fold thickness (mm)
• Insulin: 2-Hour serum insulin (mu U/ml)
• BMI: Body mass index (weight in kg/(height in m)^2)
• DiabetesPedigreeFunction: Genetic predisposition score
• Age: Age in years
• Outcome: Target variable (0 = No Diabetes, 1 = Diabetes)

The dataset is cleaned by handling missing values using median imputation.

🏆 Machine Learning Models

The pipeline implements various models to classify individuals based on their health attributes.

Model Overview

1. Decision Tree Classifier
   • A simple tree-based model that learns decision rules.
   • Good for interpretability but prone to overfitting.
2. Random Forest Classifier
   • An ensemble of decision trees, reducing variance and improving generalization.
   • Tuned using GridSearchCV to find the best hyperparameters.
3. Support Vector Machine (SVM)
   • Uses a polynomial kernel to separate classes.
   • Works well for small and medium-sized datasets.
4. k-Nearest Neighbors (KNN)
   • Classifies based on the majority vote of k nearest neighbors.
   • Performs best with well-distributed data.
5. LightGBM (Optional)
   • Gradient boosting framework for high-performance classification.
   • Used if installed; skipped otherwise.

🛠 Installation

1. Clone this repository:

   git clone https://github.com/Real-J/ml-classification.git
   cd ml-classification

2. Install dependencies:

   pip install -r requirements.txt

   If using Anaconda, install missing dependencies:

   conda install -c conda-forge imbalanced-learn lightgbm

🚀 Usage

1. Place the dataset (your_dataset.csv) in the project folder.
2. Run the script:

   python ml_pipeline.py

3. The script will:
   • Preprocess the dataset
   • Handle imbalanced data using SMOTE
   • Train multiple models
   • Tune Random Forest hyperparameters
   • Display feature importance
   • Plot the confusion matrix
   • Save the best model as best_model.pkl

📊 Model Performance Example

Original class distribution: [500 268]
Resampled class distribution: [500 500]
Accuracy with Decision Tree: 0.7450
Accuracy with Random Forest: 0.7900
Accuracy with SVM: 0.7100
Accuracy with KNN: 0.7050
Accuracy with LightGBM: 0.7800
Tuned Random Forest Accuracy: 0.81

Dataset Result

📈 Feature Importance

The script generates a feature importance plot, helping to identify the most influential features in classification.

🔍 Troubleshooting

• If you get ModuleNotFoundError, install missing libraries:

  pip install numpy pandas scikit-learn seaborn matplotlib imbalanced-learn lightgbm

• If LightGBM is not installed, it will skip training that model without affecting the rest.

🤖 Future Improvements

• Support for deep learning models (e.g., TensorFlow/Keras, PyTorch)
• AutoML integration for hyperparameter tuning
• Web API for real-time model inference

📜 License

This project is licensed under the MIT License.