🚀 Employee Attrition Prediction Model 🕵️‍♂️

Welcome to my Employee Attrition Prediction Model project repository! Predicting whether employees will stay or leave the company has never been so exciting! 🤩

📢 Project Overview

My mission is to predict employee attrition and help organizations retain their valuable talent! 💼 By harnessing the power of machine learning, I aim to make employee retention more data-driven and effective.

📈 Exploratory Data Analysis (EDA)

EDA helps uncover hidden insights in the data! Captivating count plots, mesmerizing histograms, and intriguing box plots were used to better understand the data. 📊

🧹 Data Preprocessing

Data preprocessing is the key to a successful model! Outliers were treated with magic ✨, and data was transformed using the Yeo-Johnson method. Categorical features were encoded with labels for the model's delight! 🧙‍♂️

🛠️ Feature Engineering

Some feature engineering magic was sprinkled to make the model even more powerful! 🪄 New features were created and existing ones modified to enhance predictions.

🤖 Model Development

An army of models was trained to predict employee attrition! From Logistic Regression 📈 to XGBoost 🚀, the model lineup is unstoppable!

📊 Model Evaluation

The models are as good as they come! They were evaluated using a barrage of metrics, including accuracy, precision, recall, F1-score, and ROC AUC.

🎯 Feature Selection

To make the model even more focused, Recursive Feature Elimination (RFE) was employed to choose the best features.

🔧 Hyperparameter Tuning

The models were fine-tuned using GridSearchCV to unlock their full potential!

📏 Scaling

Standard scaling is the secret weapon! It ensures that all features are on the same playing field, with mean 0 and standard deviation 1.

📂 Repository Structure

• /data: The treasure chest that holds the dataset.
• /notebooks: The magical spellbooks (Jupyter notebooks) for EDA, preprocessing, modeling, and evaluation.
• /scripts: Python scripts for data preprocessing, feature engineering, and model training.
• /models: Home to the trained model files.
• /results: Where the magic scrolls of model evaluation and visualization outputs are kept.

💻 Usage

To cast your own prediction spell and replicate the results, follow these enchanted steps:

1. 📥 Download the dataset and place it in the /data directory.
2. 🧙‍♂️ Run the notebooks and scripts in the specified order.
3. 🔮 Evaluate model performance and choose the mightiest model of them all!

---

Join me on this thrilling journey of predicting employee attrition with style and flair! Let's make employee retention great again! 🌟✨💼🔮🚀