|
svmRegression
|
The svmRegression function returns the Number Vector containing the trained
weights for the support vector machine regression on the specified training points.
The argument {X} contains the N x M independent variables in the form of a
number vector array: The argument {Y} contains the N dependent variables in the form of a number vector: The return value is a Structure {S} containing the following elements: The return element {Weights} contains the N trained weights for the support
vector machine dual form regression model in the form of a number vector: For a more thorough discussion of the complex subject of support vector machine
regression, please refer to "An Introduction to Support Vector Machines and other
kernel-based learning methods", by Nello Christianini and John Shawe-Taylor,
Cambridge University Press, 2000. When to use The svmRegression function can be used to perform linear and non-linear dual
form regressions on small, mid, and large scale training data. Training data up to
50,000 x 100 can be regressed, in reasonable time, on medium speed laptop computers.
A wide range of built-in support vector machine kernels are available plus user
defined kernel Lambdas are also readily accepted, making the svmRegression function
useful across a wide range of applications.
x11
x12
...
x1M
x21
x22
...
x2M
...
...
...
...
xN1
xN2
...
xNM
y1
y2
...
yN
w1
w2
...
wN
(svmRegression X Y kernel ETollerance maxError MaxGenerations maxSVSize printSW) The Structure {S} with elements: Error, Weights, Generations, Ey, and Py.
Here are a number of links to Lambda coding examples which contain this instruction in various use cases.
Here are the links to the data types of the function arguments. Here are also a number of links to functions having arguments with any of these data types.
You can always talk with the AIS at aiserver.sourceforge.net.
Name
Description
AIS Types X The N x M Number Vector array of independent training points Integer Y The N Number Vector of dependent training points Integer kernel The two argument support vector machine kernel function to be used
in the regression. Symbol ETollerance The error tolerance limit as a percent of dependent value. Real maxError The maximum error rate at which training may stop. Real MaxGenerations The maximum number of training generations at which training must stop. Integer MaxSVSize The maximum number of support vectors attempted during initialization. Integer printSW The verbose mode display switch for testing purposes. Symbol
Returns:
Examples
Argument Types
Vector
Structure
Integer
NumVector
Symbol
Real
How do I contact the AIS team?