Udacity-Machine Learning纳米学位－学习笔记1

课程地址

Category:

Machine Learning
   Artificial Intelligence
   Data Science
   Basic Statics

Tips

• stick to your schedule

• be relentless in searching the answer on your own

• be an active member in the community

16/5/20

Welcome

Machine Learning

What is Machine Learning ?

• Processing the data & understanding the data

• Then react intelligently to it

• Build models to represent Data

• Lots of different things, really next natural evolution

Compare with traditional programming ?

• Traditional: Build the decision making directly into the programming

• ML: Build an agent who can look at a bunch of images over time and recognize

Application ?

• Almost every field : Predict, Identify, Maximize

Objectives ?

• When to use them, How

• What to apply to solve it, How to evaluate

16/5/21

Artificial Intelligence & Data Science

2 Fields:

• Artificial Intelligence

• Data Science

Artificial Intelligence

• To create machines that are as smart as humans

• 6 Characteristics

• 5 Big problems to solve

• 4 Schools of AI

• 3 Fundamental Process of knowledge based AI

• Fundamental Tech: Bayesian Rule, Bayesian Network

Data Science

What is Data Scientist ?

• Can do math, and programming.

• Ask the right questions and solve them.

• Communicate, Report, and Present.

What does Data Scientist Do ?

• Data

• Model

• Understand patterns

Machine Learning

3 Parts:

• Supervised Learning:
  Labeled Data to get the label for new data

• Unsupervised Learning:
  Input->Observe the relationship among them->Identify

• Reinforcement Learning:
  **Learn from delayed award **

What to learn:

• Parameters

• Structure

• Hidden concepts

What for:

• Predict

• Diagnose

• Summarization

Output:

• Classification

• Regression

16/5/22

Basic Statics Concepts

Basic Statics

Measure of Central Tendency

• mode, median, average

Variability of Data

• Range= Max-Min

• Quartile: Q1, IQR=Q3-Q1

• Outlier: <Q1-1.5IQR or >Q3+1.5IQR

• Variance: average( sum( (Xi-Xbar)^2 ) )

• Standard Deviation: squared root of Variance