https://github.com/glgl0x00/bike-sharing-demand-predction-with_autogloun

https://github.com/glgl0x00/bike-sharing-demand-predction-with_autogloun

auto-gluon aws kaggle kaggle-competition kaggle-dataset python pytorch sagemaker torchvision

Last synced: 3 months ago
JSON representation

• Host: GitHub
• URL: https://github.com/glgl0x00/bike-sharing-demand-predction-with_autogloun
• Owner: GLGL0x00
• Created: 2025-01-06T00:35:12.000Z (5 months ago)
• Default Branch: main
• Last Pushed: 2025-01-06T00:39:12.000Z (5 months ago)
• Last Synced: 2025-03-15T14:12:37.952Z (3 months ago)
• Topics: auto-gluon, aws, kaggle, kaggle-competition, kaggle-dataset, python, pytorch, sagemaker, torchvision
• Language: HTML
• Homepage:
• Size: 1.49 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • Codeowners: CODEOWNERS.txt

Awesome Lists containing this project

README

        
# Bike Sharing Demand Prediction

This project is part of the Udacity and AWS Nanodegree program, focusing on predicting bike sharing demand using the Kaggle dataset "Bike Sharing Demand". The project utilizes AutoGluon with AWS SageMaker for model training and evaluation.

## Project Phases:

1. **Exploratory Data Analysis (EDA):**

   - Explored the dataset to understand its structure, distributions, and relationships between variables.

   - Identified patterns and insights to guide feature selection and engineering.

2. **Model Training with Tabular Predictor:**

   - Utilized AutoGluon's Tabular Predictor to train machine learning models on the dataset.

   - Evaluated model performance using metrics such as RMSE.

   - Generated predictions and analyzed fit_summary to understand model behavior.

3. **Hyperparameter Tuning:**

   - Tuned model hyperparameters to improve predictive performance.

## Files Included:

- `bike_sharing_demand.ipynb`: Jupyter notebook containing the code implementation.

- `report.md`: Markdown report containing information about the project work.

## Future Work:

- Explore additional feature engineering techniques to further enhance model performance.

- Experiment with different machine learning algorithms and ensembles to compare performance.

## Credits:

- Udacity for providing the Nanodegree program, access to resources, and the template for the Jupyter notebook.

- Kaggle for the "Bike Sharing Demand" dataset.

- Implementation code by Ahmed Abdelgelel.