CRANV2
Aslib Cranfield Research Project: Factors Determining the Performance of Indexing Systems: Volume 2
Appendix 5A
appendix
Cyril Cleverdon
Michael Keen
Cranfield
An investigation supported by a grant to Aslib by the National Science Foundation.
Use, reproduction, or publication, in whole or in part, is permitted for any purpose of the United States Government.
- 297 -
The typical layout would then be as shown:-
(K-l) rel. does. J LKth rel. doe.
IN-S) choices including]
R- I[OCRerr] rel. docs. J
TIME
The number of ways in which this layout can be formed is the numerator
' the required ratio and is given by (S-1)C(K_1 iN-S)C
) ' (R -K)
The denominator is simply NCR
Thus PK, S =
{S-1)C(K_I) . (N-S)C(R_K)
N
CR
QK can now be evaluated as indicated above for each value of K from
to
N (S-I)C(K_I) (N-S)c(R _K)
bus QK [OCRerr]t S . N
CR
S-1
N
(N
=[OCRerr]-" S(S -i )' -SIc(R-K)
(K-1It (S-K): NCR
6 N
= K___ [OCRerr] S' (N-S)C(R -K)
NCR K: (S-K):
S=I
N
= K CK . -SIc(R_K
NCR S=I
The terms of the series vanish for K> S> (K+N-R) so that the
mlts of S may he changed to give
K+N -R
QK = "N--cRK [OCRerr] IS(N-S)C(R_K)I Note thatCK " K[OCRerr] R
S=K