Information Technology Reference
In-Depth Information
Ta b l e 1 . 3
Notation rules
Meaning
Notation
Query
q
or
q
i
z
(i)
A quantity
z
for query
q
i
Number of training queries
n
Number of documents associated with query
q
m
Number of document pairs associated with query
q
m
˜
Feature vector of a document associated with query
q
x
m
j
Feature vectors of documents associated with query
q
x
={
x
j
}
=
1
Term frequency of query
q
in document
d
TF(q, d)
Inverse document frequency of query
q
IDF(q)
Length of document
d
LEN(d)
Hypothesis
h(
·
)
Scoring function
f(
·
)
Loss function
L(
·
)
Expected risk
R(
·
)
R(
Empirical risk
·
)
Relevance degree for document
x
j
l
j
Document
x
u
is more relevant than document
x
v
l
u
l
v
Pairwise preference between documents
x
u
and
x
v
l
u,v
Total order of document associated with the same query
π
l
Ground-truth label for document
x
j
y
j
Ground-truth label for document pair (
x
u
,
x
v
)
y
u,v
Ground-truth list for documents associate with query
q
π
y
Ground-truth permutation set for documents associate with query
q
Ω
y
π
−
1
(j)
Original document index of the
j
th element in permutation
π
Rank position of document
j
in permutation
π
π(j)
Number of classes
K
Index of class, or top positions
k
VC dimension of a function class
V
Indicator function
I
{·}
Gain function
G(
·
)
Position discount function
η(
·
)
References
1. Amento, B., Terveen, L., Hill, W.: Does authority mean quality? Predicting expert quality
ratings of web documents. In: Proceedings of the 23th Annual International ACM SIGIR
Conference on Research and Development in Information Retrieval (SIGIR 2000), pp. 296-
303 (2000)
2. Baeza-Yates, R., Ribeiro-Neto, B.: Modern Information Retrieval. Addison-Wesley, Reading
(1999)