https://github.com/miladiouss/stat-learning

https://github.com/miladiouss/stat-learning

Last synced: 3 months ago
JSON representation

• Host: GitHub
• URL: https://github.com/miladiouss/stat-learning
• Owner: Miladiouss
• Created: 2022-01-18T20:15:20.000Z (almost 4 years ago)
• Default Branch: main
• Last Pushed: 2022-01-30T10:48:39.000Z (almost 4 years ago)
• Last Synced: 2025-03-18T02:48:10.829Z (10 months ago)
• Language: Python
• Size: 2.93 KB
• Stars: 0
• Watchers: 0
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# Statistical Learning

This is a collection of my computational notes on statistical learning, machine learning, and data science. It is new and growing. The topics are oriented for experienced python programmers with a strong mathematical background who may not be familiar with the field of statistical learning.

The following points make this collection unique that you may want to consider before continuing:

1. **Plotly**: I believe Plotly is a superior data visualization library. I will not be using Matplotlib since it doesn't provide the customization I need and I'm much more productive using Plotly. On top of that, Plotly is JS based and interactive; you cannot beat that.

2. **Scipy**: Majority of machine learning tutorials out there are using sk-learn. Instead, here I will try using the stats module of Scipy since it is more fundamental, has better documentation, and is more customizable. Having said that, I will use sk-learn whenever it is superior to Scipy.

3. **Data Simulation**: Most importantly, whenever possible, I will simulate the data in the code since I believe data simulation is one of the main pillars of data science: data wrangling, data simulation, modeling, visualization.

## Terminology

$$

\bm{y'} = f(\bm{x}; \bm{\theta})

$$

- Model ($f$)

    - `param` ($\bm{\theta}$) - a.k.a: model parameters, $\bm{\Theta}$

    - Modes:

        - Training (`'train'`)

        - Evaluation (`'eval'`)

- Data

    - `x` ($\bm{x}$) - a.k.a: model input, input data, input features , $X$

    - `target` ($\bm{y}$) - a.k.a: target, target feature, y_true, $Y$

    - `output` ($\bm{y\prime}$) - a.k.a: model output, output feature, y_pred, y_true, label

- Data sets

    - Training dataset (`train_ds`)

    - Validation dataset (`valid_ds`)

    - Test dataset (`test_ds`)

- Trining

    - Loss function (`loss`)

    - Deep Learning:

        - Forward Propagation

        - Backward Propagation

        - Optimization algorithm

        - Learning rate (`lr`, $\eta$)

        - Gradient