Company Bankruptcy Prediction ML Project
Introduction
This is a machine learning project that predicts the bankruptcy of a company using various financial features. The project includes data preprocessing, model selection, and evaluation. The best-performing model has an accuracy of 96.63%.
Prerequisites
The following packages are required to run the project:
- numpy
- pandas
- scikit-learn
- matplotlib
Data
The data used in this project is obtained from a public dataset and contains financial information of companies. The features used in the project are selected based on their relevance to the bankruptcy prediction task.
Preprocessing
The data is preprocessed to handle missing values, outliers, and scaling issues. The preprocessing steps include:
- Train-test split
- Standard scaling
Models
The following machine learning models are used to predict the bankruptcy of a company:
- Logistic Regression
- K-Nearest Neighbors
- Decision Tree
- Support Vector Machine (Linear and RBF kernel)
- Neural Network
- Random Forest
- Gradient Boosting
Dimensionality Reduction
To improve the performance of the models, the number of features is reduced using PCA (Principal Component Analysis).
Results
The performance of each model is evaluated using accuracy, and the best-performing model is Support Vector Machine with an RBF kernel, achieving an accuracy of 96.63%.
Conclusion
This project provides a framework for predicting the bankruptcy of a company using machine learning algorithms. The best-performing model can be used as a reference for future projects with similar objectives.
How to run the code
- Clone the repository
- Install the required packages
- Run the bankruptcy_prediction.py file
Contributing
Contributions are welcome and appreciated. To contribute, create a pull request with your changes.
License
This project is licensed under the MIT License.