Machine Learning

Algorithms of Classic Machine Learning Problems Using MATLAB

No Machine Learning Packages used

All Self-Written Source Codes from Scratch.

Topics Included:

- Nearest Neighbor Methods (KNN Classification/Regression)

- Clustering (K-Centers, DP-Centers)

- Linear Methods:
  - LDA and Ridge Regression
  - Logistic Regression (SGD)
  - Support Vector Machine (SSGD)
  
- Dimensionality Reduction using PCA

- Kernel for SVM & Clustering

Classification

Logistic Regression & Stochastic Gradient Descent Algorithm

Training Dataset: 3 Classes, Features in R^4

SGD Algorithm Learning Progress over Iterations:

Binary-SVM with RBF Kernel

Training Dataset & Outcome Decision Boundary:

SSGD Algorithm Learning Progress over Iterations:

K-Nearest Neighbors Algorithm

Training Dataset:

Predictions By KNN Algorithm:

Clustering

K-Means Algorithm

Clustering Outcome: K = 3

WCSS Analysis:

DP-Centers Algorithm

NBA 18-19 MPG vs. PPG Dataset:

Clustering Outcome: Lambda = 44

Dimensionality Reduction

PCA

AT&T Faces Dataset: 400 images with size 112 x 92 in PGM format

Reconstructing Original Face from 'The Average Face':