Most popular machine learning algorithms implemented from scratch.
For each algorithm (except Decision Tree), the code consists of two parts:
- Manual
- Sklearn
Manual contains implementations of the algorithm from scratch, while sklearn is an implementation of the same algorithm using the sklearn library. Both parts use the same data so it is possible to compare the algorithm we have built with the one implemented in the library.
Code with 2D plots is written in the jupyter notebook, but for better visualization, the code containing 3D plots is written as a .py extension.
- Linear Regression
- Simple
- Multiple
- Polynomial
- Polynomial Multiple
- Logistic Regression
- Simple
- Multiple
- K-Nearest Neighbors
- 2D
- 3D
- K-Means Clustering
- 2D
- 3D
- Decision Tree
- Classification
- Regression
- Naive Bayes
- Classic NB
- GaussianNB
- Numpy - Package for scientific computing
- Matplotlib - Plotting library
- Seaborn - Data visualization library based on matplotlib
- Scikit-Learn - Data mining and data analysis library
- Przemysław Kanach - Przemysław Kanach