Machine Learning and Data Mining Lecture 1

Machine Learning and Data Mining Lecture 1

1. The learning problem - Outline

    1.1 Example of machine learning

　　Predicting how a viewer will rate a moive?

　　10% improvement = 1 million dollar prize

The essence of machine learning:

A pattern exists

We cannot pin it down mathematically

We have data

The following method is not machine learning.

When you tag viewer from different perspective(attributes) and predict other viewer with the similar attributes,it's not machine learning.

 Components of learning

Formalization:

　　Input: x (customer application)

Output: y (good/bad customer)

Target Function: f: x->y (ideal credit approval formula)

Data:(x₁,y₁),(x₂,y₂),(x₃,y₄),.....,(x_n,y_n)

Hypothesis:  g: x->y

 Supervised Learning

　　Example from vending machines - coin recogniztion.

The input data can be classify.

Unsupervised Learning

　　There are the data and good luck try to predict the credit.

For example, when you learning a foreign language, you have no other resource to learn , what you have is the radio . So

you listen it everyday even though you don't understand it. but eventually,your brain will build a model in your head.

when you have a teacher to teach you the foreign language, you will be able to learning that foreign language much faster.

Reinforcement Learning

we get(input, some output, grade for the output)

1.2 Components of Learning

1.3 A simple model

1.4 Types of learning

1.5 Puzzle