Geoscience Reference
In-Depth Information
APPENDIX
C
Symbols
Symbol Usage
δ
confidence parameter in PAC model
error parameter in PAC model
λ
weight
complexity of hypothesis
covariance matrix of a Gaussian distribution
θ
model parameter
μ
mean of a Gaussian distribution
incompatibility between a function and an unlabeled instance
ξ
SVM slack variable
A
an algorithm
b
SVM offset parameter
C
number of classes
c
loss function
D
observed training data
D
feature dimension; degree matrix
d
distance
F
hypothesis space
f
predictor, hypothesis (classification or regression)
H
hidden data
H
entropy
k
number of clusters, nearest neighbors, hypotheses
L
normalized graph Laplacian
L
unnormalized graph Laplacian
l
number of labeled instances
N
Gaussian distribution
n
total number of instances, both labeled and unlabeled
p
probability distribution
q
auxiliary distribution for EM
matrix transpose
u
number of unlabeled instances
W,w
graph edge weights
w
SVM parameter vector
X
instance domain
