📊 Machine Learning Experiments

Welcome to the Machine Learning Experiments repository. This repository contains various machine learning experiments exploring different techniques and algorithms across multiple datasets. Each notebook demonstrates practical applications of machine learning concepts and provides insights into different methodologies.

📂 Repository Structure

📦Machine Learning Experiments
 ┣ 📜categorical-feature-transformations.ipynb
 ┣ 📜credit-card-fraud-detection.ipynb
 ┣ 📜Stacking_Ensemble.ipynb
 ┗ 📜titanic-competition.ipynb

📝 Notebooks Overview

1. Categorical Feature Transformations

This notebook focuses on different techniques for transforming categorical features for use in machine learning models. It covers methods such as:

• Label Encoding: Converting categorical values to numerical labels.
• One-Hot Encoding: Creating binary columns for each category.
• Target Encoding: Encoding categories based on the mean target value for each category.
• Frequency Encoding: Encoding categories based on their frequency in the dataset.

Key Insights:

• Understanding various encoding techniques and their impact on model performance.
• Comparing the effectiveness of each encoding method using different datasets.

2. Credit Card Fraud Detection

This notebook addresses the problem of credit card fraud detection using various machine learning algorithms. The goal is to identify fraudulent transactions from a dataset of credit card transactions.

Techniques Utilized:

• Data Preprocessing: Handling imbalanced data using techniques like oversampling and undersampling.
• Feature Engineering: Creating new features and selecting important ones.
• Model Training and Evaluation: Implementing models such as Logistic Regression, Decision Trees, Random Forest, and XGBoost.
• Model Evaluation: Using metrics such as Accuracy, Precision, Recall, F1-Score, and ROC-AUC to evaluate model performance.

Key Insights:

• Handling imbalanced datasets effectively.
• Comparing the performance of different machine learning models for fraud detection.
• Importance of evaluation metrics in assessing model performance for imbalanced datasets.

3. Titanic Survivor Prediction (Kaggle Competition)

This notebook explores the classic Titanic survivor prediction problem. The objective is to predict whether a passenger survived based on features such as age, gender, and class.

Techniques Utilized:

• Data Preprocessing: Handling missing values and feature scaling.
• Feature Engineering: Creating new features from existing ones.
• Model Training and Evaluation: Implementing models such as Logistic Regression, Decision Trees, Random Forest, and Support Vector Machines (SVM).
• Model Evaluation: Using metrics such as Accuracy, Precision, and Recall to evaluate model performance.

Key Insights:

• Effective data preprocessing and feature engineering techniques.
• Comparing different machine learning models for binary classification.
• Insights into feature importance and model interpretability.

4. Stacking Ensemble

This notebook demonstrates the use of stacking ensemble methods to improve model performance. Stacking involves training multiple base models and combining their predictions using a meta-model.

Techniques Utilized:

• Base Models: Training various base models like Logistic Regression, Decision Trees, and Random Forest.
• Stacking: Combining the predictions of base models using a meta-model, typically a simple model like Logistic Regression.
• Model Evaluation: Assessing the performance of the stacking ensemble compared to individual base models.

Key Insights:

• Understanding the concept of stacking ensembles.
• Implementing stacking ensembles to enhance model performance.
• Comparing the effectiveness of stacking ensembles with individual models.

🛠️ Installation

1. Clone the repository:

git clone https://github.com/your-username/Machine-Learning-Experiments.git
cd Machine-Learning-Experiments

2. Create a virtual environment and activate it:

python -m venv env
source env/bin/activate  # On Windows, use `env\Scripts\activate`

3. Install the required dependencies:

pip install -r requirements.txt

🚀 Running the Notebooks

1. Open the Jupyter Notebooks:

jupyter notebook

2. Select and run the desired notebook to explore the experiments and results.

📧 Contact

For any questions or feedback, feel free to reach out to:

• Name: Aradhya Dhruv
• Email: aradhya.dhruv@example.com

📝 Acknowledgements

This repository contains experiments conducted as part of my learning and exploration in the field of machine learning. Special thanks to the creators of the datasets and the open-source community for their invaluable resources and support.

---