A theory of information with special application to search problems

Classical information theory concerns itself with communication through a n oisy channel and how much one can infer about the channel input from a know ledge of the channel output. Because the channel is noisy the input and out put are only related statistically and the rate of information transmission is a statistical concept with little meaning for the individual symbol use d in transmission. Here we develop a more intuitive notion of information t hat is concerned with asking the right questions-that is, with finding thos e questions whose answer conveys the most information. We call this confirm atory information. In the first part of the paper we develop the general th eory, show how it relates to classical information theory, and how in the s pecial case of search problems it allows us to quantify the efficacy of inf ormation transmission regarding individual events. That is, confirmatory in formation measures how well a search for items having certain observable pr operties retrieves items having some unobserved property of interest. Thus confirmatory information facilitates a useful analysis of search problems a nd contrasts with classical information theory, which quantifies the effici ency of information transmission but is indifferent to the nature of the pa rticular information being transmitted. The last part of the paper presents several examples where confirmatory information is used to quantify protei n structural properties in a search setting. (C) 2000 Elsevier Science Ltd. All rights reserved.