Semantics Based Water Information Telegraph Translation

Water information telegraph is a kind of telegraph, by which the water information observed in each water information station is transmitted obeying the rules of water information telegraph. Its head is a six-bit-character string, and its body is composed of some five-bit-character strings and expresses the water information. Blanks divide each other between the neighboring strings of a telegram. Water Information Telegraph Translation (WITT), which translates water information telegraph into water information data required by computer, is regard as a kind of Machine Translation (MT) with the aim of extracting information. The present WITT system is on the basis of rules of water information telegraph; statistically, it usually can translate about 95% of telegraph, but 4~5% is unable to be translated and is error telegraph which is out of accord with the rules and has been sent by various reasons in the process of telegraphing. At present the error telegraph is corrected by man and then is translated, in this way the information is delayed transmitting in time, so that extracting information is an obstacle to automate processing of the water information. These bring research of WITT Technology into necessity.Firstl y, the thesis presents the MT Model with aim of extracting information and its present application to WITT, which is the syntactic-analysis-based WITT based on rules of water information telegraph, and there also lies the above problem. Secondly, in order to solve the problem, the article introduces semantic analysis technique, which is a kind of language processing technique in MT and provides semantic information for translation to increase the rate of translation. Then Logic Semantics is presented in detail, which is a kind of semantic analysis technique. Thirdly, Applying Logic Semantics to WITT, this thesis constructs the logic semantic framework of telegraph to obtain the logic semantics of telegraphs. Finally, the article brings forward the semantics-based Water Information Translation Model, and then evaluates its system, and the result shows the degree of automation is improved.