Machine Learning in CSC 196K

kettledoctorΤεχνίτη Νοημοσύνη και Ρομποτική

15 Οκτ 2013 (πριν από 3 χρόνια και 8 μήνες)

80 εμφανίσεις

Machine Learning in

CSC 196K

What motivated machine
learning?

What is machine Learning?

Statistics and ML algorithms
design

Machine learning is motivated
by a class of problems


No traditional algorithmic solution


Computationally expensive


Traditional algorithm


Input => algorithm => Output

When algorithm=?, we may use machine
learning algorithm to construct a
mapping function between input and
output.

Input and Output for Machine
Learning Algorithm


For the class of ML algorithms that
generalization performance is based on
a given finite number of training
examples, we have:




Training


set


Mapping

function

Machine

Learning

algorithm

Statistics and ML algorithms
design


ML
methods overcome shortcomings of
statistical methods (estimation, classification,
prediction)


Learning bias



prior assumption of a
Bayesian


Nonlinear

problems
--

regression, linear
model


Learning systems that use a
hypothesis
space

--

hypothesis testing


Statistical learning

theory (SVM
-
Support
Vector Machine)

Types of Data mining
Techniques in CSC 196K


Techniques for data mining covered in
csc196K


Data warehousing


OLAP and data preprocessing


Statistics


Machine learning


This course only covers data mining related
ML techniques and concepts (a subset of ML
methods)


Data mining is one of many applications of
machine learning

To Learn more on Machine
Learning and Applications


CSC 219 Machine Learning


CSC 215 Artificial Intelligence


CSC 196L Intelligent Systems


Publications in CSUS Library


On
-
line info