Machine Learning, Andrew Ng

2017/09/03

1 Introduction

1.2 Welcome

Machine Learning

  • Grew out of work in AI
  • New capabilities for computers

Examples

  • Database mining
    • Large datasets from growth of automation/web.
    • E.g., Web click data, medical records, biology, engineering
  • Application cannot program by hand.
    • E.g., Autonomous helicopter, handwriting recognition, most of Natural Language Processing (NLP), Computer Vision.
  • Self-customizing programs
    • E.g., Amazon, Netflix product recommendations
  • Understanding human learning (brain, real AI).

1.3 What is machine learning

Machine Learning definition

  • Arthur Samuel (1959). Machine Learning: Field of study that gives computers the ability to learn without being explicitly programmed.
  • Tom Mitchell (1998) Well-posed Learning Problem: A computer program is said to learn from experience E with respect to some task T and some performance measure P, if its performance on T as measured by P improves with experience E.

Machine learning algorithms:

  • Supervised learning
  • Unsupervised learning

Others: Reinforcement learning, recommender systems.

Also talk about: Practice advice for applying learning algorigthms.

1.4 Supervised Learning

“right answers” given

  • Regression problem
    • Predict continuous valued output (price)
    • E.g., housing price prediction.
  • Classfication problem
    • Discrete valued output (0 or 1, or more)
    • Multiple/Infinite features might be considered, e.g., tumor size, age, clump thickness, uniformity of cell size, uniformity of cell shape
    • E.g., breast cancer (malignant, benign),

1.5 Unsupervised Learning

no “right answers” given

E.g., Google News, Organize computing clusters, Social network analysis, Market segmentation, Astronomical data analysis

Cocktail party problem

Cocktail party problem algorithm, in Octave:

[W,s,v] = svd((repmat(sum(x.*x,1),size(x,1),1).*x)*x');

2 Linear regression with one variable

2.1 Model representation

Supervised Learning: Given the “right answer” for each example in the data.

Regression Problem: Predict real-valued output.

  • Training Set => Learning Algorithm => h (hepothesis)
  • Size of house => h => Estimated price

How to represent h ? \(h_θ(x) = Θ_0 + Θ_1x\)

License: (CC 3.0) BY-NC-SA

Search

    Table of Contents