ML in Python

This is my playground repo for my refresh process of several years using matlab, and R in several ML tasks. Here, I focus on try to complete a typical stack of an ML engineer completely in Python. Classical models from the basics (trees, linear models) to the more complex ones (deep learning, nlp) will be part of those notebooks

Frameworks

1. NumPy
2. Matplotlib
3. scikit-learn
4. Tensorflow
5. CuPy and Rapids

Topics

1. Basics of NumPy (I, II), Tensors and Matplotlib
2. Web scrapping with requests and beatifulsoup
3. Decision trees
   • Random Forest (titanic example)
   • Greadient Boosted Decision trees (titanic example)
4. Regression using GLMs
   • Perfom a non-regularized logistic regression from scratch (notebook)
   • Logistic regression (titanic example)
   • Linear regression
   • Spline regression (GAMs)
5. Singular Value Decomposition and PCA
6. K-Means|K-NN
7. SVM
8. Generalized Additive Models (GAMs)
9. DBSCAN
10. Time-series forecasting
    • Heuristic
    • Machine Learning
    • Recurrent neural networks (RNNs) and long short term memory (LSTMs)
    • CNNs
11. UMAP and t-SNE
12. Deep learning
    • Regression
      • Naive classifier from scratch (FChollet)
      • Simple linear regression celsius2fahrenheit (Udacity)
    • Classification
      • Clothes classification (Udacity)
      • Dogs vs Cats (Udacity)
      • Dogs vs Cats with data augmentation (Udacity)
      • Dogs vs Cats with Transfer learning (InceptionV1) (Udacity)
      • Flowers with Transfer learning (InceptionV1) (Udacity)
      • Movie reviews (imdb) (FChollet)
      • TabPFN: Titanic dataset (Kaggle) (titanic)
13. Project: Visual search and image retrieval
14. NLP
15. Other topics: data wrangling+SQL
16. GPU acceleration

• Apple Silicon (tf.device)