JPS60163161A - Dictionary maintenance system - Google Patents

Abstract

PURPOSE:To attain accurate maintenance without requiring any detailed knowledge relating to a dictionary record by having the relation of correspondence of related dictionary constituting elements as data and comparing the data with an inputted maintenance data. CONSTITUTION:A key of a dictionary record for the purpose of maintenance is inputted from an input device 6, it is retrieved by a dictionary memory 2, and when the corresponding record exists, the record is read, divided into the constituting element units and stored in a maintenance job memory 4. Assuming that a record having a key of ''maintenance'' is stored in the maintenance job memory 4 and the change relating to Japanese word information is commanded, a series of constituting elements relating to the Japanese word information are displayed on a display device 6 together with its index. The maintenance operator uses the input device 6 to reserve the value of the constituting element written on a write enable area on the screen. After the end of maintenance, the result is collated with the content of a dictionary description rule memory 3 to check the error of data.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Application of the Invention] The present invention relates to a maintenance method for a dictionary that is a component of a system that analyzes, converts, or generates natural language.

[Background of the invention]

Conventionally, dictionaries that contain information about both languages have been used for translation between languages, but these dictionaries vary in content and headings depending on the field of the text being translated, and to improve translation performance. Maintenance such as supplementing prenatal care is essential. At that time, conventionally, maintenance confectionery makers had to be careful not to cause inconsistencies between components as a result of maintenance.
The drawback is that accurate maintenance requires detailed knowledge of the relationships between the components.

[Purpose of the invention]

In view of the above points, the present invention provides efficient and accurate maintenance that does not require detailed knowledge of dictionary records by checking not only individual errors in the data values of dictionary components but also checking inconsistencies between the components. The purpose is to provide a dictionary maintenance method that enables.

[Summary of the invention]

The above purpose is to have the correspondence between elements of the dictionary that are related to each other, such as part-of-speech codes and conjugation type codes, and to compare the data with input maintenance data. This method has been achieved.

[Embodiments of the invention]

Hereinafter, the present invention will be explained in detail using an example.

In FIG. 1, 1 is a processor, 2 is a dictionary memory, 3 is a dictionary description rule memory, 4 is a memory for maintenance work, 5 is a display device, 6 is an input device such as a keyboard, and 7 is a communication line. FIGS. 2-A and 2-B show the processing flow of the deep processor 1.

An example will be explained in which the dictionary memory 2 is for Japanese-English translation. An example of its contents is shown in FIG. Here Japanese “′
"Maintenance" is the key symbol "≠゛' followed by a number to separate the components. ÷1" is followed by the Japanese part-of-speech code, "42" is followed by the conjugation type and conjugation line code, "+3" is followed by the English part-of-speech code, and ≠4" is followed by the English continuation code. After 5'°, a three-digit number describes the form of the inflection. Dictionary description rule memory 3
The code that is allowed as a constituent element is associated with the subject and is stored in an array format. Table 1 shows the relationship between Japanese part-of-speech codes and conjugation type codes.

Table 1 These values are stored in the dictionary description rule memory 3 as 'PR,M(4
)”.In this array, “
PAR, 'l'" part of speech code, same <" INF
'S' describes the maximum number of conjugation types that the relevant part of speech can take, and the array ``KCOD (4)'' describes the conjugation type codes that the relevant part of speech can take.In the maintenance work memory 4, the constituent elements are stored in their corresponding storage areas.For the following explanation, the Japanese part-of-speech code is "f) CO
It is assumed that the usage type code is stored in "8CODE".

The dictionary maintenance method according to the present invention will be explained below with reference to FIG. 2-A.

The key of the dictionary record to be maintained is input from the input device 6. As a result, the dictionary memory 2 is searched, and if the drawing record exists (102), the record is read, divided into constituent elements, and stored in the maintenance work memory 4 (13). +1 if dictionary record does not exist
412) Initialize the maintenance work memory 4 1104).

Next, the processing content is determined by a command from the input device 6 or an instruction using a function key or the like (105 to 1
06).

It is now assumed that a record with a key of "maintenance" is stored in the maintenance work memory 4, and that a change regarding Japanese information has been instructed. Then, a series of components related to the Japanese information in the maintenance work memory 4 are displayed on the display device 6 together with headings indicating what each component represents. flr17).

An example of this state is shown in FIG. In this state, the maintenance worker only needs to use the input device 6 to maintain the values of the components written in the writable area on the screen (the area marked with a dotted line in FIG. 4). After the maintenance is completed, the input data is taken in by pressing the send key or the like (lO8), and is compared against the contents of the dictionary description rule memory 3 to check for errors in the data (+1ng). This part will be explained in more detail with reference to Figure 2-B regarding Japanese parts of speech and types of conjugation.

First, a counter for array search is initialized (201), and the currently stored part-of-speech code is checked to see if it is allowed (202). If the part-of-speech code is not allowed (203-204), it will be judged as an error +2
05). If the part of speech is allowed as a constituent element, the number of conjugation type codes that the part of speech may have is taken +2 ('16). As a result, if it is found that the part of speech is conjugated, +2 ('+7) it is checked whether the currently stored conjugation type code exists among the conjugation type codes taken by the part of speech code (208).

If it does not exist (209-210), it is determined that there is an error (211). Further, if it is determined that the part of speech is not conjugated, it is checked whether there is a currently stored conjugation type code (+212), and if so, it is judged as an error (213). In this way, if II K ul+ is input as the utilization type code in the state shown in FIG. 4, it will be determined to be an error.

+1 if an error is discovered as a result of an incorrect check
10. ) outputs an error message and +111) requests input again. If there is no error, f is stored again in the corresponding location of each component in the maintenance work memory 4.
l12). After the maintenance work is completed, the components stored in the maintenance production memory 4 are edited according to predetermined rules, fl1
3), fl14) to be stored in the dictionary memory 2.

By doing this, the maintenance contractor can perform accurate dictionary maintenance without having detailed knowledge of the constituent elements. In this example, the application is shown to check whether there is a contradiction between the part of speech and the type of conjugation, but it can also be applied to the relationships between various elements in the dictionary, such as the part of speech and the semantic code, and the part of speech and grammatical information. it is obvious.

〔Effect of the invention〕

As explained and explained above, according to the present invention, if a contradiction occurs between components as a result of maintenance, it is possible to notify the maintenance worker that there is an error, without having to have detailed knowledge of the components. This also has the effect of allowing accurate dictionary maintenance.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention, FIGS. 2-A and 2-B are flowcharts for explaining the operation of the embodiment of the present invention, and FIG. FIG. 4 is an explanatory diagram showing the storage expansion in the dictionary memory 2, and FIG. 4 is an explanatory diagram of the operation in one embodiment of the present invention. l...Processor, 2...Dictionary memory, 3...Dictionary description rule memory, 4...Memory for maintenance work, 5...
・Display device, 6... Input device, 7... Communication circuit Figure 1% Z-7) Ministry of Figure ? −β diagram

Claims (1)

[Claims] ■ In a system that handles natural language, the correspondence between words, grammar, meanings, etc. (hereinafter referred to as constituent elements) for analyzing, converting, generating, etc. natural language is described according to predetermined rules. In a method for adding, altering, or deleting the contents of the dictionary memory (hereinafter referred to as maintenance) for a device having a dictionary memory, a configuration that is a unit given a logical meaning in a dictionary record in the dictionary memory. A dictionary maintenance method whose main feature is to check for inconsistencies in the values of elements that are related to each other when maintaining elements. 2. The dictionary maintenance method according to claim 1, characterized in that when checking for inconsistencies between constituent elements, the checking is carried out while referring to a table in which dictionary description rules are written.