This project analyzes employee attrition using a neural network model to predict employee attrition and department based on various features.
File: attrition.ipynb Purpose: Predict employee attrition and department using neural networks. Dataset: Attrition Data
pandas numpy scikit-learn tensorflow Install the required libraries using pip:
bash pip install pandas numpy scikit-learn tensorflow
- Data Import and Exploration: Load and inspect the dataset.
- Data Preparation: Select features, encode categorical data, and scale numeric data.
- Model Creation: Build a neural network with shared layers and two output branches for department and attrition.
- Model Training: Train the model on the training data.
- Model Evaluation: Evaluate and print the model's accuracy.
- Metrics: Accuracy is used but may not be ideal for imbalanced data; consider precision, recall, F1 score, or AUC.
- Activation Function: Softmax is used for output layers to handle multi-class classification.
- Improvements: Feature engineering, handling class imbalance, hyperparameter tuning, regularization, exploring different architectures, ensemble methods, data augmentation, and advanced optimization algorithms.
Clone the repository. Open attrition.ipynb file Install the required libraries. Open attrition.ipynb in Jupyter Notebook. Run the cells sequentially to execute the analysis.