Machine Learning

About

Implementations and homeworks of two MOOCs courses(offered by Hsuan-Tien Lin):

• Machine Learning Foundations:
  • Corresponds to the first half-semester of the National Taiwan University (NTU) course "Machine Learning".
  • Introduce topics ranging from "When Can Machines Learn" to "Why", "How" and beyond.

Machine Learning Foundations - Course Certificate

• Machine Learning Techniques:
  • The second half-semester of the NTU course "Machine Learning".
  • Three popular tools:
    1. embedding numerous features (kernel models, such as support vector machine)
    2. combining predictive features (aggregation models, such as adaptive boosting)
    3. distilling hidden features (extraction models, such as deep learning).

Machine Learning Techniques - Course Certificate

courses are based on the textbook Learning from Data: A Short Course.

Syllabus

Machine Learning Foundations

When Can Machines Learn?

• The Learning Problem | Learning to Answer Yes/No
• Types of Learning | Feasibility of Learning

Why Can Machines Learn?

• Training versus Testing | Theory of Generalization
• The VC Dimension | Noise and Error

How Can Machines Learn?

• Linear Regression | Linear 'Soft' Classification
• Linear Classification beyond Yes/No | Nonlinear Transformation

How Can Machines Learn Better?

• Hazard of Overfitting | Preventing Overfitting I: Regularization
• Preventing Overfitting II: Validation | Three Learning Principles

Machine Learning Techniques

Embedding Numerous Features

• Linear Support Vector Machine | Dual Support Vector Machine
• Kernel Support Vector Machine | Soft-Margin Support Vector Machine
• Kernel Logistic Regression | Support Vector Regression

Combining Predictive Features

• Bootstrap Aggregation | Adaptive Boosting
• Decision Tree | Random Forest
• Gradient Boosted Decision Tree

Distilling Hidden Features

• Neural Network | Deep Learning
• Radial Basis Function Network | Matrix Factorization

Content

• Kernel SVM & Soft-Margin SVM
• Kernel Logistic Regression and Support Vector Regression
• Blending and Bagging
• Adaptive Boosting
• Decision Tree and Random Forest
• Gradient Boosted Decision Tree
• kMeans, k-Nearest Neighbors
• Radial Basis Function Network
• Neural Network and Deep Learning
• Autoencoder

Usage

Clone the repo and use the virtualenv:

git clone https://github.com/AaronYALai/Machine_Learning_Techniques

cd Machine_Learning_Techniques

virtualenv venv

source venv/bin/activate

Install all dependencies and run the model:

pip install -r requirements.txt

cd Autoencoder

python autoencoder.py