Hello! Welcome to my project on predicting breast cancer survival. In this notebook, I've tackled the challenge of forecasting survival outcomes in breast cancer patients using machine learning. This work is not only a demonstration of my skills in data science but also contributes to the important field of medical prognosis.
The primary goal of this project is to develop a predictive model using clinical data and gene expression profiles. By doing so, I aim to enhance the understanding and accuracy of survival predictions in breast cancer care.
The data for this project comes from the "Breast Cancer Gene Expression Profiles (METABRIC)" dataset. It consists of:
- Clinical information, including patient details and death status.
- Gene expression profiles.
- Gene mutation data (though this part is not utilized in the analysis).
I started by splitting the dataset into three distinct parts:
- Clinical dataset
- Gene expressions dataset
- Gene mutation dataset (not used in further steps)
A thorough EDA was performed for each dataset. For instance, in the clinical dataset, I focused on the 'overall_survival' variable and removed features like 'death_from_cancer' to avoid redundancy.
I carefully engineered features from the clinical data and gene expressions to make them more suitable for modeling and to enhance the predictive power of the machine learning algorithms.
I experimented with various machine learning models. I have implemented a neural network model from scratch for clinical, gene expressions and reduced gene expressions datasets. My approach involved rigorous testing and validation to ensure the reliability and accuracy of the predictions.
The models showed promising results in predicting breast cancer survival. I have included a detailed analysis of the model performance, backed by metrics and visualizations, in the notebook.
This project extensively uses Python and several of its libraries, such as:
- Pandas for data manipulation
- Scikit-learn for modeling
- Matplotlib/Seaborn for visualizations
To interact with this project:
- Clone the repository containing this Jupyter Notebook.
- Ensure you have Python installed along with the necessary libraries.
- Run the notebook in a Jupyter environment, executing cells sequentially to see the workflow and results.
Thank you for visiting my project! I hope it offers valuable insights into the application of machine learning in healthcare and inspires further research in this vital area.