Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/sadegh15khedry/titanic-survival-prediction
Predicting Titanic survival using neural networks.
https://github.com/sadegh15khedry/titanic-survival-prediction
conda-environment jupyter-notebook kersa matplotlib numpy pandas python seaborn sklea supervised-learning tensorflow titanic-kaggle
Last synced: 3 days ago
JSON representation
Predicting Titanic survival using neural networks.
- Host: GitHub
- URL: https://github.com/sadegh15khedry/titanic-survival-prediction
- Owner: sadegh15khedry
- License: mit
- Created: 2024-07-07T02:24:33.000Z (4 months ago)
- Default Branch: master
- Last Pushed: 2024-08-10T12:50:58.000Z (3 months ago)
- Last Synced: 2024-08-11T11:59:17.393Z (3 months ago)
- Topics: conda-environment, jupyter-notebook, kersa, matplotlib, numpy, pandas, python, seaborn, sklea, supervised-learning, tensorflow, titanic-kaggle
- Language: Jupyter Notebook
- Homepage:
- Size: 5.59 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
## Overview
This repository contains a machine learning project for predicting survival on the Titanic using a neural network.
## Table of Contents
- [Project Overview](#project-overview)
- [Dataset](#dataset)
- [Project Structure](#project-structure)
- [Setup Instructions](#setup-instructions)
- [Data Preprocessing](#data-preprocessing)
- [Model Training](#model-training)
- [Model Evaluation](#model-evaluation)
- [Results](#results)
- [Future Work](#future-work)
- [Contributing](#contributing)
- [License](#license)
## Project Overview
The goal of this project is to predict whether a passenger survived the Titanic disaster based on various features such as age, gender, class, etc. We employ various data preprocessing techniques and build an Artificial Neural Network (ANN) to make these predictions.
## Dataset
The dataset used in this project is the Titanic dataset available from [Kaggle](https://www.kaggle.com/c/titanic/data). It contains information about the passengers including whether they survived or not.
## Project Structure
The project has the following structure:
```
titanic-survival-prediction/
│
├── datasets/
│ ├── raw/
│ │ └── titanic.csv
│ └── read/
│ ├── train/
│ │ ├── X_train.csv
│ │ └── y_train.csv
│ ├── val/
│ │ ├── X_val.csv
│ │ └── y_val.csv
│ └── test/
│ ├── X_test.csv
│ └── y_test.csv
│
├── notebooks/
│ ├── data_exploration.ipynb
│ ├── data_preprocessing.ipynb
│ ├── model_training.ipynb
│ └── model_evaluation.ipynb
│
├── src/
│ ├── data_preprocessing.py
│ ├── model_training.py
│ ├── model_evaluation.py
│ └── utils.py
│
├── results/
│ ├── model_accuracy.png
│ ├── model_loss.png
│
├── README.md
└── environment.yml
```
## Setup Instructions
1. **Clone the repository:**
```bash
git clone https://github.com/sadegh15khedry/Titanic-Survival-Prediction
cd Titanic-Survival-Prediction
```
2. **Create and activate a conda environment:**
```bash
conda env create -f environment.yml
conda activate titanic
```
3. **Download the dataset:**
- Download the dataset from the [Kaggle Titanic competition](https://www.kaggle.com/c/titanic/data).
- Place the downloaded files in the `datasets/raw/` directory.
- Run the preprocessing script to split the data into train, validation, and test sets and place them in the respective folders in `datasets/ready/`.
## Data Preprocessing
Data preprocessing involves handling missing values, encoding categorical variables, and normalizing numerical features. The preprocessing steps can be found in the `data_preprocessing.py` script and are also detailed in the `data_preprocessing.ipynb` notebook.
## Model Training
We use an Artificial Neural Network (ANN) to predict survival. The network architecture, training procedure, and evaluation are implemented in `model_training.py` and `model_training.ipynb`.
## Model Evaluation
The model's performance is evaluated on the validation set. Metrics such as accuracy, precision, recall, and F1-score are used to assess the model. Evaluation results and visualizations can be found in `model_evaluation.py` and `model_evaluation.ipynb`.
## Results
### Training
- Train Loss: 0.3781
- Train Accuracy: 0.8296
### Validation
- Val Loss: 0.5708
- Val Accuracy: 0.8146
*Train and validation loss and accuracy*
### Testing
- Test accuracy: 0.754
- Test Loss: 0.669
- Precision: 0.753
- recall: 0.754
- F1: 0.749
```
precision recall f1-score support
0 0.76 0.86 0.80 105
1 0.75 0.61 0.67 74
accuracy 0.75 179
macro avg 0.75 0.73 0.74 179
weighted avg 0.75 0.75 0.75 179
```
*Confusion Matrix*
## Future Work
- **Hyperparameter Tuning:** Experiment with different architectures, learning rates, and other hyperparameters.
- **Ensemble Methods:** Combine predictions from multiple models to improve accuracy.
## Contributing
Contributions are welcome! Please open an issue or submit a pull request for any enhancements or bug fixes.
## License
This project is licensed under the MIT License. See the `LICENSE` file for more details.