Đang chuẩn bị nút TẢI XUỐNG, xin hãy chờ
Tải xuống
Components of the machine-readable dictionary can be applied in a number of information systems. The most direct applications of the kind are in wordprocessing or in "writingsupport" systems built on a wordprocessing base. However, because a central function of any dictionary is in fact data verification, there are other proposed applications in communications and data storage and retrieval systems. Moreover, the complete interrelational electronic dictionary is in some sense the model of the language; and there are, accordingly, additional implications for language-based information search and retrieval. In regard to wordprocessing, the electronic lexicon can serve as the base for. | MACHINE-READABLE COMPONENTS IN A VARIETY OF INFORMATION-SYSTEM APPLICATIONS Howard R. Webber Reference Publishing Division Houghton-Mifflin Company 2 Park street Boston. MA 02108 Components of the machine-readable dictionary can be applied in a number of information systems. The most direct applications of the kind are in wordprocessing or in writingsupport systems built on a wordprocessing base. However because a central function of any dictionary is in fact data verification there are other proposed applications in communications and data storage and retrieval systems. Moreover the complete interrelational electronic dictionary is in some sense the model of the language and there are accordingly additional implications for language-based information search and retrieval. In regard to wordprocessing the electronic lexicon can serve as the base for spelling verification in which the computer detects many spelling or typographical errors and spelling correction in which the computer offers corrections to the errors it has identified . Because it is possible to develop algorithms that permit the computer to calculate the chances that the single best alternative it offers is actually correct this substitution can in many cases be made automatically. It is at this point in the development of such systems wise to flag such automatic corrections for inspection by the operator. At the present time these processes generally depend upon the application of strict frequency measures which permit the lexicon to be reduced to small-machine proportions and thereby reduce the possibility of a false hit the passing of a misspelled common word that happens to coincide in orthography with a legitimate but rare word. As our ability to draw cognitive information from text increases and as available memory increases then such limits can be abandoned. Truncation of the lexicon for other specific applications can be considered. It is possible for example to shape the lexicon to reflect a
