JPH0439109B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION <Industrial Application Field> The present invention relates to a translation device that performs analysis, conversion, and generation processing on an input sentence and outputs a translated sentence.

<Prior Art> In a conventional machine translation device, if a parenthesis symbol is included in an input sentence, it cannot be translated as is. For this reason, the only option was for the operator to delete the parentheses and translate them one by one.

However, such a method imposes a heavy burden on the operator, and it is extremely difficult to obtain the desired translation result (translated text containing parentheses at predetermined positions).

<Object of the Invention> The present invention has been made in view of the above-mentioned conventional problems, and an object of the present invention is to provide a translation device that can process even when parentheses are included in an input sentence.

<Example> The present invention will be explained in detail below based on the drawings.

FIG. 1 is a block diagram of a translation device according to the present invention.

This translation device includes a central processing unit (CPU) 1 that performs processing such as translation processing, a main memory 2 that stores programs, etc. for processing by the CPU 1, and a main memory 2 that displays information necessary for the translation processing. A display device (CRT) 3, a keyboard 4 for inputting information necessary for the translation process, a translation module 5 for storing information necessary for the translation process, and a translation dictionary, grammar rules, and structure conversion rule table 6. It is being

The above-mentioned translation module 5 is for translating a source language (input sentence) into a target language (translated sentence), as will be described later.
It includes buffers A, B, C, D, and E, a CPU, and translation program memory. These buffers A to E store data at each analysis level in the translation process, as will be described later.

FIG. 3 shows the analysis level of automatic translation performed by translation module 5.

Now, when the source language is input from the upper left, dictionary lookup, morphological analysis, and syntax analysis and analysis proceed. Machine translation can be broadly divided into two types depending on the level of analysis. One is the pivot method, which analyzes concepts that do not depend on each language (called an intermediate language) and generates a target language from there. The other method is called the transfer method, which includes morphological analysis, syntactic analysis,
The internal structure of the source language is obtained by performing either semantic analysis or context analysis, and then the target language is generated after converting it to the same level structure of the target language.

Here, each analysis will be explained as follows.

・ Dictionary lookup, morphological analysis - Look up the dictionary shown in Figure 1, 6, and check the grammatical information such as part of speech for each word.
Obtain translations and analyze tense, person, number, etc.

・ Syntactic analysis - Determines the structure of a sentence, such as dependencies between words.

・Semantic analysis - Distinguish what is semantically correct from the results of multiple syntactic analyzes and what is not.

- Contextual analysis - understand the topic and eliminate omissions and ambiguity.

It is assumed that the above translation module has been performed up to at least the level of parsing. That is, as shown in FIG. 2, it consists of four parts: a dictionary lookup morphological analysis section 1, a syntactic analysis section 2, a language conversion section 3, and a target language generation section 4.

FIG. 5 shows the content of each buffer in FIG. 4 when the English sentence ``This is a pen.'' is translated into Japanese.

First, the original text is read into buffer A in FIG. 4 as shown in FIG. 5a. Necessary information for each word is retrieved by the dictionary retrieval unit shown in FIG. 2 and placed in buffer B shown in FIG. The part of speech information that is part of it is
The result will be as shown in Figure 5b. This is a multi-part speech word, which is uniquely determined by the syntactic analysis unit shown in FIG. 2, and the dependency relationship is entered into buffer C in FIG. 4 as shown in FIG. 5 c. At this time, Fig. 1, 6
From the grammar rules (also shown in Figure 4), sentence → subject, predicate subject → noun phrase predicate → verb, noun phrase noun phrase → pronoun noun phrase → article, noun and analyze. This rule expresses that ``A sentence is made up of a subject and a predicate.'' (first rule). In the conversion section of FIG. 2, analysis is performed according to the rules of tree structure conversion, similar to syntax analysis, and the results are entered into the buffer D of FIG. 4 as shown in FIG. 5d. In the generation section of FIG. 2, appropriate particles and auxiliary verbs are added, the Japanese form is formed, and the result is placed in buffer E of FIG. 4 as shown in FIG. 5e. This is the output of the translation module.

Next, the processing operation of the present invention will be explained using FIGS. 6 to 8.

First, in the dictionary lookup process (Figure 6), the pointer is set to the first word (n1), and it is checked whether the word is in parentheses, (,,).If it is not in parentheses, the normal dictionary lookup process ( n3) is performed. If the word is a parenthesis, count the number of words up to the closing parenthesis corresponding to that word, i.e., ), and
Add to the word count of the previous word. This information is
It is included in the buffer of dictionary lookup results shown in Figure b. As a result, in the syntax analysis process, the number of words in a pair of parentheses 2, the number of words in the part surrounded by parentheses, and the number of words in the word before the parentheses are summed up and looked up in a dictionary. Since the content of the resulting buffer is entered in the column for the number of words before the parentheses, the part enclosed in parentheses can be parsed in the same way as the other parts. The above processing is performed for all words, and the dictionary lookup processing is completed.

After the dictionary lookup process, the syntax analysis and conversion process are performed as usual, and in the subsequent generation process, the part enclosed in parentheses is generated.

As shown in FIG. 7, when the generation process for one word (n1) is completed, the translated word is stored in the buffer (n2). At this time, check whether the next word is parentheses, (,) (n3). If it is not parentheses,
Normal generation processing continues. If the word is a parenthesis, the words up to the closing parenthesis corresponding to the parenthesis, ie, ), are stored in the buffer (n4). The above process is repeated and the generation process is completed.

FIG. 8 shows, as an example, disk (40
An example of the output when , disk (40 Mbytes), is input is shown.

<Effect> As explained in detail above, the translation device of the present invention translates parentheses included in an input sentence,
Since it is equipped with a means for outputting it as is at a predetermined position in a translated sentence, it is possible to process parentheses that appear at any position in an input sentence. For this reason, conventionally it was necessary for the operator to delete the parentheses in the input sentence, but according to the present invention, such operations are no longer necessary, and the usability of the apparatus can be significantly improved.

[Brief explanation of drawings]

Figure 1 is a block configuration diagram of the device of the present invention, Figure 2 is a diagram showing a schematic configuration example of a translation module, Figure 3 is a diagram showing analysis levels, Figure 4 is a detailed diagram of the translation module, and Figure 5 a. 6 to 8 are diagrams showing the contents of each buffer in FIG. 4, FIG. 6 is a flowchart of dictionary lookup processing, FIG. 7 is a flowchart of generation processing, and FIG. 8 is a diagram showing a display example. 1 is CPU, 2 is main memory, 3 is CRT, 4
is the keyboard, and 5 is the translation module.

Claims (1)

[Claims] 1. In a translation device that performs translation processing of analysis, conversion, and generation of an input sentence and outputs a corresponding translated sentence, when a parenthesis symbol is included at a certain position in the input sentence from the beginning. , although the input sentence containing the string surrounded by parentheses is subjected to the translation process,
A means for parsing the entire input sentence by treating the character string within the parentheses together with the word immediately before the character string as one word; The string is parsed and translated, and
A translation device comprising: means for not translating the character string surrounded by parentheses, but outputting the character string of the input sentence as is at a position corresponding to the position in the translated sentence.