© 1978 by British Computer Society
| ||||||||||||||||||||||||||||||||||||||||||||||||||
The relationship between entropy and file search times
Computation Department, UMIST, P.O. Box 88, Manchester, UK
The formula for entropy, used in information theory, can be reinterpreted as the selectivity of a structured index for a retrieval system. This paper shows that for one indexing method, for both successful an unsuccessful searches, there is a direct relationship between the saving in comparisons, if the index structure is used, and the entropy of the grouping of retrieval keys in the index. If the number of keys is large compared with the number of subfiles in the index, the result has a very simple form.
Received October 1976.
* Computation Department, UMIST, P.O. Box 88, Manchester M60 1QD