AUTO ML APP: Developing and Deploying Dash via Heroku

Creating an AUTO-ML app with Dash. Includes classification and time series alorithms. The app implements an explained-ML philosophy. Classification includes dimensionality reduction with t-sne for visualization and PCA. Classification uses simple Logistic Regression and Support Linear Vector. Time series include autocorrelation and XGBoost Regression. Includes basic visual explanations from all algorithms.

Web app: https://auto-ml-app.herokuapp.com/

Classification Algorithms

Forecasting Algorithms

Credits

• DASH (https://plotly.com/dash/)
• PYMC3 (https://docs.pymc.io/)
• Wattenberg, et al., "How to Use t-SNE Effectively", Distill, 2016 (https://distill.pub/2016/misread-tsne/)
• Understanding LSTM Networks (http://colah.github.io/posts/2015-08-Understanding-LSTMs/), (http://blog.echen.me/2017/05/30/exploring-lstms/)
• T. Chen, C. Guestrin, "XGBoost: A Scalable Tree Boosting System" , arXiv:1603.02754v3 (https://arxiv.org/abs/1603.02754v3)
• Christopher Olah for inspiration (http://colah.github.io/posts/2015-09-Visual-Information/)