JPS6398774A - Language analyzing device - Google Patents

Abstract

PURPOSE:To analyze a proper noun into plural types of meaning in accordance with the context by using a property information adding means to give the property information to the character strings which are not recorded into a dictionary means and the character strings whose property information are not recorded into the dictionary means among those input character strings. CONSTITUTION:An input processing part 14 reads out the stored input sentences and outputs them to a unit segmenting part 16. A dictionary retrieving part 22 retrieves a reference dictionary 20 for the input sentences sent from the part 16 and divided into dictionary consulting units. A character string if serves as a proper noun is outputted to a proper noun processing part 24; while the character strings which are not equal to proper nouns are outputted to a processing part 36 and stored in a dictionary information storing table 36a of the part 36. A default property information adding part 34 is read out of the table 36a and sent to the part 22 through the part 16. The property information is added to the proper nouns having no property information according to the retrieving result of the dictionary 20 obtained by the part 22. These proper nouns are stored in the table 36a. A syntax structure analyzing part 38 analyzes the syntax structure of an input sentence that is read out of the table 36a and undergone the morpheme analysis.

Description

TECHNICAL FIELD The present invention relates to a language analysis device, particularly to a language analysis device useful for automatic translation devices.

Prior Art For example, when creating a corresponding Japanese sentence from a sentence in a foreign language such as English, the human-generated form 7U elements of the English sentence are analyzed, the syntax is analyzed, the sentence structure is converted, and then the sentence structure is converted. Generate a Japanese translation.

That is, by searching a dictionary, the form jEi elements of each application forming the input sentence are analyzed, and information such as parts of speech for these 16 forms is obtained. after that.

Based on the obtained information such as parts of speech, the structure of the input sentence is analyzed by analyzing the modifying relationships between each word or the blocks formed by these words using grammatical rules. , converts the sentence structure of the input sentence into the order of the Japanese syntactic arrangement based on the parsed syntax, generates Japanese morphemes according to the converted arrangement, and generates a Japanese sentence.

In the morphological analysis performed when generating a translated sentence, the input sentence is divided into dictionary lookup units using delimiters such as spaces, commas, and colons, and the dictionary is searched for these dictionary lookup units to identify morphemes such as words. The part of speech and other information are obtained. In this case, for example, some proper nouns can have two or more different meanings depending on the context.

For example, r(lsaka C1tyJ is rOsaka
There are cases where it is used to represent a group, such as ``C1ty decided...'', and cases where it is used to indicate a group, such as ``...in 0saka...''.
There are cases where it is used to indicate a place, such as C1ty J. Traditionally, only one meaning of a proper noun was stored in a dictionary, so it was not possible to handle this type of usage.
Analysis accuracy was low.

OBJECTS OF THE INVENTION It is an object of the present invention to provide a language analysis device capable of solving the drawbacks of the prior art and analyzing a proper noun into multiple meanings depending on the context.

Structure In order to achieve the above object, the present invention includes an input means for manually inputting a character string in a predetermined language, a dictionary means used for searching the character string input from the input means, and a dictionary means for searching the input character string. As a result of the search by the search means, character strings that are not recorded in the dictionary means and character strings for which character information is not recorded in the dictionary means are sealed with property information. The fourth property information attaching means is characterized in that a plurality of property information is attached to a character string without property information. Hereinafter, a detailed explanation will be given based on one embodiment of the present invention.

Figure 1 shows the language analysis equipment according to the present invention. l1
An example of application to a No. 11 translation device is shown. In addition,
The present invention can be used for purposes other than morphological analysis of English input sentences, and can be used not only for English-Japanese translation devices that translate English into Japanese, but also for automatic translation devices that translate one language into another language. Needless to say, it can be applied effectively.

This embodiment has an input processing section 14, into which data is input from the input device 12. Input table 2t12
includes, for example, a keyboard having character keys such as alphanumeric keys, function keys, etc., an optical character reading device for reading alphabetic text recorded on paper, a magnetic disk reading device, and the like.

The input processing unit 14 has an input character string buffer 14a, and has an input table! ! The English input sentence input from 12 is stored in the input character string buffer 14a. The input processing section 14 reads the input sentence stored in the input character string buffer 14a and outputs it to the unit extraction section 1B. The unit extraction unit 16 is a functional unit that performs dictionary lookup unit extraction of the input sentence sent from the input processing unit 14 by searching the delimiter table 18. The delimiter table 18 stores delimiters such as spaces and commas.

The unit extraction unit IB reads delimiters from the delimiter table 18, and divides the input sentence sent from the input processing unit 14 into character strings that serve as units when searching the reference dictionary 20 by dividing the input sentence at the portion where the delimiter exists. 0
The divided character strings are input to the dictionary search section 22.

The dictionary search section 22 searches the reference dictionary 20 for the input sentence divided into dictionary lookup units sent from the unit extraction section 16 . In the reference dictionary 20, for example, as shown in FIG. 2, entries of character strings of syntax, their parts of speech, property information, etc. are recorded. In addition to the proper nouns shown in the figure, the reference dictionary 20 also records character strings of other parts of speech, such as verbs and adjectives. In addition, in the figure, the record of a proper noun as a part of speech refers to the registered proper noun processing described later, and does not represent a normal grammatical proper noun. indicates what it represents, and is not limited to one. This is because, as will be described later, a proper noun represents a plurality of properties depending on the case in which it is used.

The dictionary search unit 22 searches the reference dictionary 2G for the character string divided into dictionary lookup units, and if the character string is a proper noun, outputs it to the proper noun processing unit 24 to perform proper noun processing to be described later. In addition, if it is not a proper noun, the processing unit 3
6 and the dictionary information storage table 38a of the processing unit 36.
Save to.

The proper noun processing unit 24 includes the immediately preceding sentence-final processing unit 2B,
j It consists of a proper noun processing section 28, a proper noun itself processing section 30, a previous proper noun and proper noun processing section 32, and a default property information adding section 34.

The immediately preceding sentence end processing unit 28 determines whether the character string immediately preceding the character string searched and manually entered by the dictionary search unit 22 is the end of the sentence, and if the immediately preceding character string is the end of the sentence, the character string to be processed is processed. Dictionary search unit 2 converts the first uppercase letter of the column to lowercase letter
2, and causes the dictionary search unit 22 to search the reference dictionary 20 again. Character strings that are not found in this second search are determined to be unregistered proper nouns, and are sent to the processing unit 3B and stored in the dictionary information storage table 36a. If the immediately preceding character string is not at the end of a sentence, it is sent to the processing unit 3B as a proper noun with unknown property information and registered in the dictionary information storage table 38a, as will be described later.

The immediately preceding proper noun processing section 28 analyzes the property information of the immediately preceding character string sent from the immediately preceding sentence end processing section 2B, and outputs it to the proper noun itself processing section 30. The proper noun itself processing unit 30 examines the property information of the proper noun to be analyzed, and as described later,
If property information is not registered for either the proper noun or the immediately preceding proper noun, the proper noun and the immediately preceding proper noun are analyzed together using the registered property information of the other, and stored in the dictionary information storage table 38a of processing g2I3B. save.

The immediately preceding proper noun and proper noun processing 8II32 checks the common part of the property information of the proper name tone to be analyzed and the immediately preceding proper noun sent from the proper noun itself processing unit 30, and analyzes these proper nouns based on the common part. , is output to the processing unit 36 and stored in the dictionary-1 information storage table 38a of the processing unit 3B.

The default property information adding section 34 is read out from the dictionary information storage table 38a, sent to the dictionary search section 22 through the unit extraction section 16, and is sent to the dictionary search section 22 by the dictionary search section 22.
As a result of searching for 0, for proper nouns without property information,
Add property information. It is effective for proper nouns to have various properties depending on the situation in which they are used, so all necessary property information should be added. For example, "person, place, organization, etc." Granted. After adding property information to the proper noun, the default property information adding unit 34 sends this data to the processing unit 36 and stores it in the dictionary information storage table 38a of the processing unit 36.

The processing unit 3B has a dictionary information storage table 36a, and after storing data sent from the immediately preceding proper noun and proper noun processing unit 32, the default property information adding unit 34, or the dictionary search unit 22 in the dictionary information storage table 38a. , reads the saved data and outputs it to the syntax analysis unit 38. The syntactic analysis unit 38 performs syntactic analysis of the morphologically analyzed input sentence read from the dictionary information storage table 38a.

The operation of this device will be explained with reference to the flowchart shown in FIG.

First, an English input sentence is input from the input device 12 to the input processing unit 14.
(100). The input sentence read into the input processing unit 14 is stored in the input character string buffer 14a. The input sentence stored in the input character string buffer 14a is read out by the unit extraction section 16.

When the input sentence is input, the unit extraction section 16 reads delimiters from the primitive table 18 and performs extraction in dictionary lookup units (102).

That is, by separating the character strings constituting the input sentence with delimiters such as spaces and colons, the beginning of the character string can be added to the search key string, which is the tti position when searching the reference dictionary 20. It is divided in order from It is determined whether or not the divided dictionary lookup unit, that is, the search key string, is the end (104). If there is still a search key string (not the end), the search key string is sent to the dictionary search unit 22.

When the search key string is sent to the dictionary search unit 22, the dictionary search unit 22 searches the reference dictionary 20 for this search key string (10B).The reference dictionary 2 as shown in FIG.
108) to determine whether the entry 0 has the search key character Yushu (108), and if there is an entry, refer to the reference dictionary 20.
The part-of-speech information stored in is read out, and it is determined whether the test key character string is a proper noun (110).

If the search key character string is not a proper noun, the dictionary search unit 22 uses the data read from the reference dictionary 20 in the processing unit 3B.
and records it in the dictionary information storage table 38a (11
2) The data is saved in the dictionary information storage table 38a.
Then, the processing section 36 inputs an input to the effect that the data has been saved and the data of the examination key character string that was saved immediately before.

As a result, the process returns to step 102 and the unit cutout f
In lllB, dictionary lookup units are extracted.

If the search key character string is a proper noun in step 110, the dictionary search unit 22 reads the data of the proper noun (hereinafter simply referred to as proper noun) from the reference dictionary 20,
It is sent from the dictionary information storage table 38a of the processing unit 38 to the proper noun processing unit 24 along with the data of the previous search key character string inputted to the dictionary search unit 22 through the unit extraction unit 16, and is registered in the dictionary in the proper noun processing unit 24. Proper noun processing is performed (+14).

Here, the processing of proper nouns registered in the dictionary will be explained using the flowchart shown in FIG.

The data sent from the dictionary search section 22 to the proper noun processing section 24 is input to the immediately preceding proper noun processing section 28 through the immediately preceding sentence end processing section 26. In processing registered proper nouns, the immediately preceding sentence end processing unit 26 does not function.

The immediately preceding proper noun processing unit 28 determines whether the search key character string immediately preceding the proper noun is a proper noun that has not been registered in the reference dictionary 20, that is, has undergone dictionary-unregistered proper noun processing to be described later. If the proper noun is an unregistered proper noun, the proper noun and the entire previous unregistered proper noun are judged to be proper nouns that have proper noun property information202), and these data are processed by the processing unit. Send to 36,
It is recorded in the dictionary information storage table 36a (218).

In the immediately preceding proper noun processing unit 28, in step 200,
If it is determined that the search key string immediately before the proper noun is not an unregistered proper noun, it is determined whether the search key string immediately before the proper noun is a proper noun registered in the reference dictionary 20. (204). If the search key string immediately before a proper noun is a registered proper noun, the property information of the proper noun immediately before it is unknown.
That is, it is determined whether or not it is not registered in the reference dictionary 20 (20B).

If the property information of the immediately preceding proper noun is unknown, the process proceeds to step 202, and the proper noun and the entire immediately preceding proper noun are determined to be proper nouns having proper noun property information (202).
, the immediately preceding proper noun processing unit 28 sends these data to the processing unit 3B. The data sent to the processing unit 36 is recorded in the dictionary information storage table 3ea (218).

If the immediately preceding proper noun processing unit 28 determines in step 20B that the property information of the immediately preceding proper noun is not unknown, that is, it is registered in the reference dictionary 20, these data are subjected to the immediately preceding proper noun processing. Part 28
The proper noun itself is sent to the proper noun processing unit 30. In proper noun self processing F'A30, property information of the proper noun is unknown (
(208)
, when the property information of a proper noun is unknown, the proper noun itself processing unit 30 determines that the proper noun and the immediately preceding proper noun are all proper nouns that have the property information of the immediately preceding proper noun (210). The data is sent to the processing section 36. The data sent to the processing unit 36 is stored in the information storage table 3.
8a (218).

When the proper noun itself processing unit 30 determines that the property information of the proper noun itself is not unknown, that is, it is registered in the reference dictionary 20, the proper noun [1 body processing f1!
30 sends these data to the immediately preceding proper noun and proper noun processing section 32 . Immediately preceding proper noun and proper noun processing unit 3
In step 2, it is determined whether or not there is a commonality of property information between the proper noun and the immediately preceding proper noun (212), and if there is a commonality, the entire proper noun and the immediately preceding proper noun are
214)
, these data are sent to the processing section 3B. The data sent to the processing unit 36 is recorded in the dictionary information storage table Ha (218).

If there is no common property information between the proper noun and the immediately preceding proper noun, the proper noun is determined to be a proper noun that has property information retrieved from the reference dictionary 20, in addition to the proper noun directly 1i1. 21B) and sends the data to the processing section 36. The data sent to the processing unit 36 is recorded in the dictionary information storage table 38a (218.l).

Returning to FIG. 3, in step 10, the reference dictionary 20
If there is no search key string in the entry, determine whether the first character of the search key string is an uppercase letter (116).
, if it is not a capital letter, the dictionary search unit 22 determines that this search key character string is an unregistered word, sends it to the processing unit 38, and records it in the dictionary information storage table 38a (118).

If the first character is a capital letter, the data of this search key string is sent from the dictionary search section 22 to the proper noun processing section 24 along with the data of the previous search key string, and proper nouns not registered in the dictionary are processed. It is done (120).

The processing of proper nouns not registered in the dictionary will be explained with reference to FIG.

The data of the search key string is sent to the immediately preceding sentence end processing section 26 along with the data of the immediately preceding search key string, and the immediately preceding sentence end processing section 26 determines whether the end of the immediately preceding search key string is an end of sentence candidate ( 300).

To determine whether this is a candidate for the end of a sentence, use the previous search key character
This is done by determining whether the end of is a candidate for the end of a sentence, such as a period (,) for tundoku.

If the last word Jll= of the immediately preceding search key character string is a candidate for the last sentence, the data is sent from the immediately preceding sentence end processing section 2B to the immediately preceding proper name, +4 processing section 28, and the immediately preceding proper noun processing section 28 uses the immediately preceding search key Convert the character string to the end of the sentence and r4 (302), convert the uppercase letters of the search key character string [1] to lowercase letters, and send it to the dictionary search unit 22.

The dictionary search unit 22 searches the reference dictionary 20 for the search key character string converted to lowercase letters (304), the reference dictionary 20
It is determined whether there is an entry in (30B). If there is an entry, the dictionary search unit 22 sends the data searched from the reference dictionary 20 to the processing unit 36 and records it in the dictionary information storage table 3Eta (308); if there is no entry, the dictionary search unit 22 The processing unit 3 returns the first character of this search key string to uppercase and treats it as an unregistered proper noun.
6 and recorded in the dictionary information storage table 38a (3
10).

In step 300, if the immediately preceding sentence end processing unit 26 determines that the end of the immediately preceding search key character string is not a sentence ending candidate, the immediately preceding sentence ending processing unit 26 sends the immediately preceding proper noun processing unit 28
The data is sent to , and the immediately preceding proper noun processing unit 28 determines that the immediately preceding search key character string is not at the end of the sentence (312).
, from the immediately preceding proper noun processing section 28 to the proper noun itself processing section 30
The data is sent to , and the proper noun processing unit 30 determines that the search key character string is a proper noun whose property information is unknown (314
).

The proper noun itself processing section 30 returns the data to the immediately preceding proper noun processing section 28, where the proper noun registered in the dictionary is processed (31B). This processing of the proper noun registered in the dictionary is shown in FIG. Same as shown.

Returning to FIG. 3, if the dictionary lookup unit cut out in step 104 is the end, the dictionary search unit 22 uses the default property information material! A signal to that effect is sent to the j part 34, and the default property information material q part 34 reads out the information recorded in the dictionary information storage table 36a of the processing part 36, and adds default property information to the proper noun (122 )
.

The operation of adding default property information to a proper noun will be explained with reference to FIG.

In the default property information appended section 34, a pointer is first set at the beginning of the data read from the dictionary information storage table 36a (400). Set the pointer to the search key string at the beginning of the input sentence to which information has been added. It is determined whether the search key character string indicated by the pointer is a proper noun (402), and if it is a proper noun, the property information of the proper noun is unknown (402).
n) is determined (404), and if it is not a proper noun, the process proceeds to step 408, and the pointer is advanced to the next search key character string.

If the property information of the proper noun is unknown in step 404, default property information is added (4
0B) When attaching default property information, default property information is added to a proper noun whose property information is unknown, as shown in the lower part of FIG. For example, as shown in the several pages of the same figure, all property information such as "one person, place, group, etc." is added to the proper noun r Johnson J, whose property information is unknown. By attaching all property information to a proper noun whose property information is unknown in this way, we leave room for the proper noun to be analyzed into any of multiple properties in later parsing. be able to.

If the property information of the proper noun is not unknown in step 404, the process proceeds to step 40B, where the pointer is advanced to the next search key character string.

It is determined whether the search key string indicated by the pointer is the end (408), and if it is not the end, the process returns to step 402 and the next search key string is determined as 'I' whether it is a proper noun or a platform.
I cut it off. If the search key character string indicated by the pointer is at the end, the addition of default information ends γ.

After the default property information has been assigned to the proper noun,
The data is recorded in the dictionary information storage table 36a. After that, process af! 3B to dictionary information storage table 38a
124 to output the data recorded in the syntax analysis unit 38.
), the morphological analysis of this example ends.

The operation of this device explained above will be explained using an example of an input sentence.

Input sentence r In Tokyo St as shown in Figure 7
Mr. ation, Waiter met John
A case where onJ is input will be explained. First, human processing (100) reads this input sentence into the human processing unit 14.
The next step is to do the following, and the town i pull unit! /J Exit 102
) and divides the input sentence ";" into each word by spaces. First, the reference dictionary 20 is searched for rInJ (10B), since there is an entry for "rIn" in the reference dictionary 20 (108) and it is not a proper noun (10B).
10) Record the data retrieved from the reference dictionary 20 in the dictionary information storage table 38a (112).

Next, the reference dictionary 20 is searched for r Tokyo J (108), r Tokyo J is searched in the reference dictionary 2
There is no entry in 0 (10B) and the first character is a capital letter (11B).

Processing proper nouns not registered in the dictionary (120) Proceeds to Figure 5, since the immediately preceding one is r'InJ and there is no candidate for the end of the sentence.
(300), rInJ should be recognized as not the end of the sentence312)
, r Tokyo 4 with unknown property information (Unknown
) as a proper noun (314), and processes the proper noun registered in the dictionary (31B).

Proceed to Figure 7JS4, and the immediately preceding r InJ is not an unregistered proper noun ((200), nor is it a registered proper noun (
204), a proper name having its own property information alone, that is, a proper name with unknown property information (lJnknown) is recorded as a single unit (216).

Returning to the 7tS3 diagram, next the reference dictionary 20 is searched for rstation J (10B), rstati
on J has an entry in the reference dictionary 20 (108)
, Because it is a proper noun (110), Process the proper noun registered in the dictionary (114) Proceed to the Er54 diagram, and because the immediately preceding r Tok7a J is an unregistered proper noun (
200), r Tokyo.

Record the entire 5tation J as a proper noun with the property information "location, organization" of rStation J (
202).

Next, the reference dictionary 20 in FIG. 3 is searched for rMr, J (10B), rMr, J is the reference dictionary 20.
Since there is an entry in and it is a proper noun (' I O)
, Processing 4H registered proper nouns (tia) ,
Proceeding to Figure 4, the immediately preceding rstation J is not an unregistered proper noun (200), but a registered proper noun (
204), the property information “location/group” is not unknown (
20S, rMr, J has the property information “person” and is not unknown (208), so the immediately preceding rstati
It is determined whether or not there is a common portion of property information between on J, rMr, and J (212).

r 5tation J is "place, organization", r Mr, J
is a "person" and there is no common part between them, so rMr,
J is registered alone as a proper noun of the property information "person" (21B).

Next, return to the fIS3 diagram again and search the reference dictionary 20 for rWalterJ (10B), rWalt
erJ has an entry in the reference dictionary 20 (ios),
Since it is a proper noun (110), the proper noun registered in the dictionary is processed (114). Proceed to FIG.
, " is not an unregistered proper noun (200), but a registered proper noun (204), and the property information is "person" and is not unknown (20B), and the property information of rWalteB is "
Since "persons, places, and groups" are not unknown (208), there is a "common part of property information" (212), so "Ni r, Walter
J is collectively recorded as a proper noun with the property information "person" (214).

Next, I searched the reference dictionary 20 for "■et J, and there was an entry (108), and it was not a proper noun (11).
0), the data retrieved from the reference dictionary 20 is recorded in the dictionary information storage table 38a (112).

Furthermore, the reference dictionary 20 is searched for rJohnson J (10B), and since there is no entry for rJohnson J (108), and the first character is a capital letter (IIEI), proper nouns not registered in the dictionary are processed (120). , Proceeding to Figure 5, since there is no candidate for the end of the sentence for the immediately preceding rBetJ (300), rmetJ is recognized as not the end of the sentence (312), and r Johnson J is recognized as a proper noun with unknown property information (31).
4) Processes proper nouns registered in the dictionary (31Ei).

Proceeding to Figure 4, the previous rBetJ is not an unregistered proper noun (200), nor is it a registered proper noun (200).
04), r JahnSon'' is recorded as a proper noun with unknown property information.

After the above processing, default property information is added to the proper noun shown in FIG.

Set the pointer to rInJ, which is the middle of the dictionary lookup at the beginning (400), because it is not a proper noun (402)
, advance the pointer (408) , r Tokyo
Set to J. rTokyo J is a proper noun (402), and the nature information indicates that the entire rTokyo 5tation J is a place and group in the process of the registered proper name j4, and there is no unknown (404), so move the pointer forward ( 408), r
5tation J.

rstation J rMr, J is also a proper noun (402), and the property information is not unknown (
404), advance the pointer (40B), [Zen et J
Since is not a proper noun (402), the pointer is advanced (408).

r Jobnson J is a proper noun (402)
Since the property information is unknown (404), default property information is added (40B), as shown in FIG.

r Johnson J. Characteristic information [person, place, organization,
Others].

As described above, according to this embodiment, an English input sentence is divided into search key character strings, which are first searched using the reference dictionary 20,
Registered proper noun processing is performed when there is an entry as a proper noun in the reference dictionary 20.The registered proper noun processing takes into consideration the immediately preceding search key character string 1 and determines that the immediately preceding search key string is a proper noun. In this case, check the property information of the previous search key string and target proper noun, and if there is no property information for either, add the property information of the other, and if both have property information, remove the common part. This is used as property information of these proper nouns. Therefore, it is possible to appropriately add property information to a proper name tone without property information, and to narrow down the added property information to more appropriate property information.

This allows for effective parsing and appropriate translation in subsequent syntactic analysis.

Further, regarding a character string that is not registered in the reference dictionary 20, if the first character is a capital letter and the facing character string is determined to be the end of a sentence, the uppercase character is converted to a lowercase character and the reference dictionary 20 is searched again. Also refer to dictionary 20 for character strings at the beginning of sentences.
You can search by. Furthermore, if a character string that starts with a capital letter appears in a place that is not the beginning of a sentence, it is determined to be a proper noun, and the property information of this proper noun will be added if there is a proper noun with property information registered before or after it. Ru. Therefore, reference 2i1! It is possible to analyze proper nouns that are not registered in Uruguay 20 to some extent.

Furthermore, all necessary property information is attached to proper nouns that do not have property information attached, and unnecessary pieces of property information are removed during processing of the word, so that property information is not included in advance. It is also possible to analyze unknown or unregistered proper nouns.

In addition, multiple pieces of property information are assigned to a specific proper noun, and the appropriate property information is selected based on the property information of the preceding and following proper nouns, so when analyzing a proper noun that has multiple types of property information. Appropriate property information can be selected according to the context, and input sentences can be analyzed effectively.

Effects According to the present invention, multiple pieces of property information of proper nouns are stored in the dictionary, and multiple pieces of property information are assigned to proper nouns for which property information is not registered in the dictionary, so that proper nouns can be used. It is possible to obtain appropriate property information depending on the situation and effectively analyze input sentences.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the language analysis device according to the present invention, FIG. 2 is a diagram showing an example of data stored in the reference dictionary of FIG. 1, and FIG. A flowchart showing the overall operation; Figure 4 is a flowchart showing the processing of proper nouns registered in the dictionary among the operations shown in Figure 3; Figure 5 is a flowchart showing the processing of proper nouns not registered in the dictionary among the operations shown in Figure 3. FIG. 6 is a flowchart showing sealing processing of default property information among the operations shown in FIG. 3. FIG. 7 is a diagram showing an example in which an input sentence is processed in the apparatus shown in FIG. 1 and data stored in a dictionary information storage table changes. 1 14, ... Input processing section 1B, ... Unit extraction section 18, ... Delimiter table 20, ... Reference dictionary 22, ... Dictionary search section 24, ... Noun processing unit 2B, Immediately preceding sentence-end processing unit 28, Immediately preceding proper noun processing unit 30, Proper noun itself processing unit 32, Immediately preceding proper noun and proper noun processing unit 34,...
, default property information adding unit 313, , processing unit

Claims (1)

[Claims] 1. An input means for inputting a character string in a predetermined language; a dictionary means used for searching the character string input from the input means; and a dictionary means for searching the input character string. a search means for searching, and as a result of the search by the search means, among the input character strings, character strings that are not recorded in the dictionary means and character strings for which property information is not recorded in the dictionary means;
A language analysis device comprising: property information adding means for adding property information, wherein the property information adding means adds a plurality of property information to the character string without property information. 2. The device according to claim 1, further comprising: unit cutting means for dividing the input character string into units for searching the dictionary means; A language analysis device comprising: analysis means for analyzing the character string together with the immediately preceding character string after the character string is searched by the search means. 3. In the apparatus according to claim 2, when either the character string or the immediately preceding character string has property information and the other character does not have property information, the analysis means determines whether the property information is absent. A language analysis device characterized in that a character string is given property information of another character string. 4. In the device according to claim 2, when the character string and the immediately preceding character string both have property information, the analyzing means may extract common property information from the character string and the immediately preceding character string. A language analysis device characterized in that the property information of 5. The language analysis device according to any one of claims 2 to 4, wherein the character string analyzed by the analysis means and the immediately preceding character string are proper nouns. . 6. In the device according to claim 2, when the character string starts with an uppercase letter and the immediately preceding character string is at the end of a sentence, the analyzing means converts the first uppercase letter of the character string to a lowercase letter. After that, the language analysis device is characterized in that the search means searches the dictionary means, and if the result of the search is that the character string is not registered in the dictionary means, the character string is analyzed as an unregistered proper noun.