Upload
edan
View
49
Download
2
Tags:
Embed Size (px)
DESCRIPTION
Fuzzy Support Vector Machines (FSVM s ). Weijia Wang , Huanren Zhang , Vijendra Purohit , Aditi Gupta. Outline. Review of SVMs Formalization of FSVMs Training algorithm for FSVMs Noisy distribution model Determination of heuristic function Experiment results. SVM – brief review. - PowerPoint PPT Presentation
Citation preview
Fuzzy Support Vector Machines (FSVMs)
Weijia Wang, Huanren Zhang, Vijendra Purohit, Aditi Gupta
Outline
• Review of SVMs
• Formalization of FSVMs
•Training algorithm for FSVMs• Noisy distribution model
• Determination of heuristic function
• Experiment results
SVM – brief review
• Classification technique
• Method:
• Maps points into high-dimensional feature space
• Finds a separating hyperplane that maximizes the margin
Set S of labeled training points:
Each point belongs to one of the two classes,
Let be feature space vector, with mapping from to feature space
Then equation of hyperplane:
For linearly separable data, Optimization problem:
Subject to
For non-linearly separable data (soft margin), introduce slack variables
Optimization problem:
-> some measure of amount of misclassifications
Limitation: All training points are treated equal
FSVM – Fuzzy SVM
• each training point belongs exactly to no more than one class
• some training points are more important than others- these meaningful data points must be classified correctly
(even if some noisy, less important points, are misclassified).
Fuzzy membership: si
: how much point xi belongs to one class (amount of meaningful information in the data point)
: amount of noise in the data point
Set S of labeled training points:
Optimization problem:
large C -> narrower margin, less misclassifications
- Regularization constant
Lagrange function:
Taking derivatives:
Optimization problem:
Kuhn-Tucker conditions :
λ – lagrange multiplier
g(x) – inequality constraint
Points with are support vectors (lie on red boundary).
=> Points with same could be different types of support vectors in FSVM due to
=> SVM – one free parameter (C)
FSVM - number of free params = C, si (~ number of training points)
lies on margin of hyperplane
Two types of support vectors:
misclassified if > 1
Training algorithm for FSVMs
Objective function for optimization
Minimization of the error function Maximization of the margin The balance is controlled by tuning C
Selection of error function
Least absolute value in SVMs
Least square value in LS-SVMs Suykens and Vanewalle, 1999 the QP is transformed to solving a linear system the support values are mostly nonzero
Selection of error function
maximum likelihood method when the underlying error probability can be estimated optimization problem becomes
Maximum likelihood error
limitation
the precision of estimation of hyperplane depends on estimation of error function
the estimation of error is reliable only when the underlying hyperplane is well estimated
Selection of error function
Weighted least absolute value each data is associated with a cost or
importance factor
when the noise distribution model of data given px(x) is the probability that point x is not a noise optimization becomes
Weighted least absolute value
Relation with FSVMs take px(x) as a fuzzy membership, i.e
px(x) = s
Selection of max margin term
Generalized optimal plane (GOP)
Robust linear programming(RLP)
Implementation of NDM
Goal build a probability distribution model for data
Ingredients
a heuristic function: highly relevant to px(x) confident factor: hC trashy factor: hT
Density function for data
Density function for data
Heuristic function
Kernel-target alignment
K-nearest neighbors
Basic idea: Outliers have higher probability to be noise
Kernel-target alignment
Measurement of how likely the point xi is noise.
K-nearest neighbors: example
Gaussian kernel
can be written as
the cosine of the angel between two vectors in the feature space
The outlier data point xi will have smaller value of fK(xi,yi)
Use fK(x,y) as a heuristic function h(x)
K-nearest neighbors (k-NN)
For each xi, the set Sik
consists k nearest neighbors of xi
ni is the number of data points in the set Si
k that the class label is the same as the class label of data point xi
Heuristic function h(xi)=ni
Comparison of two heuristic function
Kernel-target alignment Operate in the feature space, use the
information of all data points to determine the heuristic for one point
k-NN Operate in the original space, use the
information of k data points to determine the heuristic for one point
How about combine them two?!
Overall Procedure for FSVMs
1. Use SVM algorithm to get the optimal kernel parameters and the regularization parameter C
2. Fix the kernel parameters and the regularization parameter C, determine heuristic function h(x), and use exhaustive search to choose the confident factor hc and trashy factor hT, mapping degree d and the fuzzy membership lower bound σ
kiS
Experiments
Data with time property
SVM results for data with time property
FSVM results for data with time property
Experiments
Two classes with different weighting
Results from SVM
Results from FSVM
Experiments
Using class center to reduce effect of outliers.
Results from SVM
Results from FSVM
Experiments (setting fuzzy membership)
Kernel Target Alignment
Two step strategy
Fix fUBk and fLB
k as following:
fUBk = maxi fk (xi, yi) and fLB
k = mini fk (xi, yi)
Find σ and d using a two-dimensional search.
Now, find fUBk and fLB
k
Experiments (setting fuzzy membership)
k-Nearest Neighbor
Perform a two-dimensional search for
parameters σ and k.
kUB = k/2 and d=1 are fixed.
Experiments
Comparison of results from KTA and k-NN with other classifiers (Test Errors)
Conclusion
FSVMs work well when the average training error is high, which means it can improve performance of SVMs for noisy data.
No. of free parameters for FSVMs is very high C, si for each data point.
Results using KTA and k-NN are similar but KTA is more complicated and takes more time to find optimal values of parameters.
This papers studies FSVMs only for two classes, multi-class scenarios are not explored.